Change feed
Every pricing change and breaking deprecation we've detected, newest first.
Type
Topic
-
Tracking since Jun 25, 2026 · 30 snapshots on file
Arena Elo (text, overall) updated.
Full history & current snapshot →View raw diff +26 −26
- amazon-nova-experimental-chat (26-02-10) 1448 - Claude Fable 5.1 (Max) 1512 - Claude Opus 5 (Max) 1509 - ERNIE-5.0-Preview (1203) 1442 - Gemini 3.7 Flash (High) 1489 - Gemini 3.8 Flash (High) 1499 - Gemini 4 Argon (High) 1532 - GLM-5.3-Flash 1471 - GLM-5V-Turbo 1437 - GPT-5.5 (High) 1471 - GPT-6 Astra (Max) 1441 - GPT-6.1 Sol (Max) 1443 - Grok 4.20 Beta 1 1444 - Grok 4.20 Multi-Agent Beta (0309) 1450 - Grok 4.6 (High) 1429 - Kimi K3 (Max) 1476 - MiMo-V2.6-Flash 1453 - MiniMax-M3 1433 - Mistral Large 3 1429 - Mistral Large 4 1431 - Muse Spark 1.3 (Max) 1490 - Muse Wonder (Max) 1502 - Qwen3.5-397B-A17B 1437 - Qwen3.7-Plus 1454 - Qwen3.8-Max 1483 - Step 5 Preview 1440 + amazon-nova-experimental-chat (26-02-10) 1449 + Claude Fable 5.1 (Max) 1513 + Claude Opus 5 (Max) 1508 + ERNIE-5.0-Preview (1203) 1443 + Gemini 3.7 Flash (High) 1490 + Gemini 3.8 Flash (High) 1498 + Gemini 4 Argon (High) 1526 + GLM-5.3-Flash 1472 + GLM-5V-Turbo 1438 + GPT-5.5 (High) 1472 + GPT-6 Astra (Max) 1442 + GPT-6.1 Sol (Max) 1444 + Grok 4.20 Beta 1 1445 + Grok 4.20 Multi-Agent Beta (0309) 1451 + Grok 4.6 (High) 1430 + Kimi K3 (Max) 1475 + MiMo-V2.6-Flash 1452 + MiniMax-M3 1432 + Mistral Large 3 1428 + Mistral Large 4 1432 + Muse Spark 1.3 (Max) 1491 + Muse Wonder (Max) 1501 + Qwen3.5-397B-A17B 1438 + Qwen3.7-Plus 1455 + Qwen3.8-Max 1484 + Step 5 Preview 1442
-
Tracking since Jun 25, 2026 · 2 snapshots on file
MiniMax-M3.1-Flash model added with license:other and availability date of 2026-10-17
- MiniMaxAI/MiniMax-M3.1-Flash
- license:other
- 2026-10-17
Full history & current snapshot →View raw diff +1 −0
+ MiniMaxAI/MiniMax-M3.1-Flash license:other 2026-10-17 -
NVIDIA Nemotron models changed
Tracking since Jun 25, 2026 · 40 snapshots on file
nvidia/Flora-Captioner license type changed from unknown to other.
- license field changed from? to other
Full history & current snapshot →View raw diff +1 −1
- nvidia/Flora-Captioner license:? 2026-10-01 + nvidia/Flora-Captioner license:other 2026-10-01
-
Tracking since Jun 25, 2026 · 66 snapshots on file
Documentation clarified that Claude Haiku 3.5 is 'retired, except on Claude Models on Google Cloud (Google managed)' instead of just 'retired, except on Google Cloud'.
Full history & current snapshot →View raw diff +1 −1
- | Claude Haiku 3.5 (retired, except on Google Cloud ) | 1,000 | 100,000 4 | 20,000 | + | Claude Haiku 3.5 (retired, except on Claude Models on Google Cloud (Google managed) ) | 1,000 | 100,000 4 | 20,000 |
-
Mistral models & deprecations changed
Tracking since Jun 25, 2026 · 18 snapshots on file
Added Z.ai GLM 5.3 Flash model with 1M-token context window for agentic workflows and visual coding.
- Z.ai GLM 5.3 Flash
- 1M-token context window
- v5.3-flash
Full history & current snapshot →View raw diff +5 −0
+ Z.ai GLM 5.3 Flash + A third-party multimodal model from Z.ai with a 1M-token context window, optimized for agentic workflows and visual coding. + Z.ai GLM 5.3 Flash + A third-party multimodal model from Z.ai with a 1M-token context window, optimized for agentic workflows and visual coding. + v5.3-flash
-
Anthropic model deprecations changed
Tracking since Jun 25, 2026 · 50 snapshots on file
Documentation clarified that Claude Models on Google Cloud (Anthropic managed) is now an Anthropic-operated platform, while Google-managed variants are partner-operated with separate retirement schedules.
- Added Claude Models on Google Cloud (Anthropic managed) as Anthropic-operated platform
- Separated Google Cloud into two categories: Anthropic managed and Google managed partner-operated
- Updated reference links to include Claude Models on Google Cloud documentation pages
Full history & current snapshot →View raw diff +1 −1
- The dates on this page apply to Anthropic-operated platforms: the Claude API, Claude Platform on AWS, and Microsoft Foundry. Partner-operated platforms (Amazon Bedrock and Google Cloud) set their own retirement schedules, so a model's lifecycle status and dates can differ. See the Amazon Bedrock, Amazon Bedrock (Opus 4.6 and earlier), and Google Cloud model tables. + The dates on this page apply to Anthropic-operated platforms: the Claude API, Claude Platform on AWS, Claude Models on Google Cloud (Anthropic managed), and Microsoft Foundry. Partner-operated platforms, Amazon Bedrock and Claude Models on Google Cloud (Google managed), set their own retirement schedules, so a model's lifecycle status and dates can differ. See the Amazon Bedrock, Amazon Bedrock (Opus 4.6 and earlier), and Claude Models on Google Cloud (Google managed) model tables.
-
Mistral AI reported benchmarks updated
Tracking since Jun 25, 2026 · 17 snapshots on file
Mistral Large 4: 12 benchmark claims (via web search)
Full history & current snapshot → -
Google Gemini API pricing changed
Tracking since Jun 25, 2026 · 30 snapshots on file
Veo 3.1 video models deprecated and removed from pricing; multiple Gemini models with future price increases scheduled for 2026-2027.
Veo 3.1 models (Standard, Fast, Lite) and related video generation features deprecated and removed from pricing; will shutdown October 22, 2026Full history & current snapshot →View raw diff +0 −12
- Billing is based on total output token consumption, calculated at a rate of 5,792 tokens per second of 720p video. Under Standard pricing, this equates to an effective price of approximately $0.10 per second. - Warning: Veo 3.1 Preview (veo-3.1-generate-preview, veo-3.1-fast-generate-preview, and veo-3.1-lite-generate-preview) is deprecated and will be shut down on October 22, 2026; migrate to Gemini Omni Flash (gemini-omni-1.1-flash) or the GA models available through the Gemini Enterprise Agent Platform to avoid service disruption. - Veo 3.1 Standard video with audio price (default) - $0.40 (720p and 1080p) - $0.60 (4k) - Veo 3.1 Fast video with audio price (default) - $0.10 (720p) - $0.12 (1080p) - $0.30 (4k) - Veo 3.1 Lite video with audio price (default) - $0.05 (720p) - $0.08 (1080p)
-
Tracking since Jun 25, 2026 · 313 snapshots on file
OpenAI shifted pricing page focus from API token rates to ChatGPT Business/Enterprise subscription plans
OpenAI API pricing page removed—ChatGPT Business/Enterprise subscription plans now featured insteadFull history & current snapshot →View raw diff +214 −42
- Price - Input:$10.00 / 1M tokens Cached input:$1.00 / 1M tokens Output:$50.00 / 1M tokens - Price - Input:$2.00 / 1M tokens Cached input:$0.10 / 1M tokens Output:$10.00 / 1M tokens - Price - Input:$0.10 / 1M tokens Cached input:$0.01 / 1M tokens Output:$0.50 / 1M tokens - Price - Image:$8.00 / 1M tokens for inputs$2.00 / 1M tokens for cached inputs$30.00 / 1M tokens for outputs Text:$5.00 / 1M tokens for inputs$1.25 / 1M tokens for cached inputs - Price - Image:$8.00 / 1M tokens for inputs$2.00 / 1M tokens for cached inputs$30.00 / 1M tokens for outputs Text:$5.00 / 1M tokens for inputs$1.25 / 1M tokens for cached inputs - Price - $0.05 per minute / $0.00083 per second - Price - Audio:$32.00 / 1M tokens for inputs$0.40 / 1M tokens for cached inputs$64.00 / 1M tokens for outputs Text:$4.00 / 1M tokens for inputs$0.40 / 1M tokens for cached inputs$24.00 / 1M tokens for outputs Image:$5.00 / 1M tokens for inputs$0.50 / 1M tokens for cached inputs - Price - Audio:$10.00 / 1M tokens for inputs$0.30 / 1M tokens for cached inputs$20.00 / 1M tokens for outputs Text:$0.60 / 1M tokens for inputs$0.06 / 1M tokens for cached inputs$2.40 / 1M tokens for outputs Image:$0.80 / 1M tokens for inputs$0.08 / 1M tokens for cached inputs - Price - $0.017 per minute / $0.00028 per second - Price - $0.0045 per minute / $0.00008 per second - Price - $0.034 per minute / $0.00057 per second - Price - $10.00 / 1k calls Search content tokens are free. - Price - Now:1 GB for $0.03 / 64GB for $1.92 per container Starting March 31, 2026:1 GB for $0.03 / 64GB for $1.92 per 20-minute session per container - Provides lower costs for requests in exchange for slower response times and occasional resource unavailability. Ideal for non-production or lower priority tasks. - Enterprise offerings - Contact our sales team to learn more about Data residency(opens in a new window), Scale Tier andReserved Capacity designed for cutting-edge customers running larger workloads. - We recommend experimenting with all of these models in the Playground(opens in a new window) to explore which models provide the best price performance trade-off for your usage. - Do you offer an enterprise package or SLAs? - We offer different tiers of access to our enterprise customers that include SLAs, lower latency, and more. Please contact our sales team to learn more. - Yes, we treat Playground usage the same as regular API usage. You will be billed at the per-token input and output prices mentioned above. - How will I know how many tokens I’ve used each month? - A token is a mathematical representation of natural language. Log in to your account to view your usage tracking dashboard(opens in a new window). This dashboard will show you how many tokens you’ve used during the current and past billing cycles. - You can set a monthly budget in your billing settings(opens in a new window), after which we’ll stop serving your requests. There may be a delay in enforcing the limit, and you are responsible for any overage incurred. You can also configure an email notification threshold to receive an email alert once you cross that threshold each month. We recommend checking your usage tracking dashboard(opens in a new window) regularly to monitor your spend. - Is access to the API included in ChatGPT Plus, Business, Enterprise or Edu? - No, OpenAI APIs are billed separately from ChatGPT Plus, Business, Enterprise and Edu. ChatGPT subscription pricing can be found at openai.com/chatgpt/pricing/. - Images are converted into tokens and charged per token. Text models price image tokens at standard text token rates, while GPT Image and gpt-realtime uses a separate image token rate. Models like gpt-4.1-mini, gpt-4.1-nano, and o4-mini convert images into tokens differently. Learn more in our docs(opens in a new window). - ChatGPT Business(opens in a new window) - ChatGPT Enterprise(opens in a new window) - ChatGPT for Education(opens in a new window) + Image 1 + Business + A secure workspace with company context and flexible seat types for any budget. + Get started(opens in a new window) + Standard seat + $20/ month + $20/month if billed annually. $25/month if billed monthly. + Premium seat + $100/ month + 5x more usage than standard, with no 5-hour limit + $100/month if billed annually. $125/month if billed monthly. + No training on your business data by default + Mix and match seat types + For teams of 2–200 employees. Unlimited subject to abuse guardrails. Learn more(opens in a new window) + Image 2 + Enterprise + Enterprise-grade AI, security, and support for businesses operating at scale + Contact our sales team to discuss enterprise pricing.*
-
Tracking since Jun 25, 2026 · 66 snapshots on file
Minor formatting change: 'Usage' page reference bolded for emphasis; no numeric limit changes.
Full history & current snapshot →View raw diff +1 −1
- In addition to providing token and request charts, the Usage page provides two separate rate limit charts. Use these charts to see what headroom you have to grow, identify when you might be hitting peak use, understand what rate limits to request, and learn how to improve your caching rates. The charts visualize a number of metrics for a given rate limit (for example, for each model): + In addition to providing token and request charts, the Usage page provides two separate rate limit charts. Use these charts to see what headroom you have to grow, identify when you might be hitting peak use, understand what rate limits to request, and learn how to improve your caching rates. The charts visualize a number of metrics for a given rate limit (for example, for each model):
-
GPQA Diamond (Epoch) scores changed
Tracking since Jun 25, 2026 · 35 snapshots on file
GPQA Diamond accuracy (0–1) updated.
Full history & current snapshot →View raw diff +1 −0
+ deepseek-v4.1-flash-1017_max 0.932 -
Tracking since Jun 25, 2026 · 7 snapshots on file
Removed model allenai/intent-aware-lfqa-qwen3-4b-multiview with odc-by license and expiration date 2026-04-13.
- allenai/intent-aware-lfqa-qwen3-4b-multiview
- license:odc-by
- 2026-04-13
Full history & current snapshot →View raw diff +0 −1
- allenai/intent-aware-lfqa-qwen3-4b-multiview license:odc-by 2026-04-13 -
Tracking since Jun 25, 2026 · 66 snapshots on file
Minor wording refinements to rate limits documentation; no numeric limit changes.
Full history & current snapshot →View raw diff +12 −12
- The API enforces service-configured limits at the organization level, but you may also set user-configurable limits for your organization's workspaces. - When usage reaches a spend limit you set, requests return HTTP 400 with error type invalid_request_error. The message begins You have reached your specified API usage limits, or You have reached your specified workspace API usage limits for a workspace limit, and states when access resumes. Raise or remove the limit to restore access sooner. - The rate limits for the Messages API are measured in requests per minute (RPM), input tokens per minute (ITPM), and output tokens per minute (OTPM) for each model class. If you exceed any of the rate limits you will get a 429 error describing which rate limit was exceeded. When the response has a retry-after header, wait that many seconds before retrying. - Many API providers use a combined "tokens per minute" (TPM) limit that may include all tokens, both cached and uncached, input and output. For most Claude models, only uncached input tokens count toward your ITPM rate limits. This is a key advantage that makes the rate limits effectively higher than they might initially appear. - Example: With a 2,000,000 ITPM limit and an 80% cache hit rate, you could effectively process 10,000,000 total input tokens per minute (2M uncached + 8M cached), because cached tokens don't count toward your rate limit. - To make the most of your rate limits, cache repeated content such as system instructions and prompts, large context documents, tool definitions, and conversation history; see prompt caching for guidance. With effective caching, you can substantially increase your actual throughput without raising your rate limits. Monitor your cache hit rate on the Usage page in the Claude Console to tune your caching strategy. - The Message Batches API has its own set of rate limits which are shared across all models. These include a requests per minute (RPM) limit to all API endpoints and a limit on the number of batch requests that can be in the processing queue at the same time. A "batch request" here refers to part of a Message Batch. You may create a Message Batch containing thousands of batch requests, each of which count toward this limit. A batch request is considered part of the processing queue when it has yet to be successfully processed by the model. - Claude Managed Agents endpoints are rate-limited per organization. These limits are separate from the Messages API rate limits above. - In addition to providing token and request charts, the Usage page provides two separate rate limit charts. Use these charts to see what headroom you have to grow, identify when you may be hitting peak use, understand what rate limits to request, and learn how to improve your caching rates. The charts visualize a number of metrics for a given rate limit (for example, per model): - Setting lower limits for Workspaces - To protect Workspaces in your Organization from potential overuse, you can set custom spend and rate limits per Workspace. - Workspace limits are set per limiter type (such as requests per minute, input tokens per minute, or output tokens per minute). + The API enforces service-configured limits at the organization level, but you can also set user-configurable limits for your organization's workspaces. + When usage reaches a spend limit you set, requests return HTTP 400 with error type invalid_request_error. The message starts with You have reached your specified API usage limits, or You have reached your specified workspace API usage limits for a workspace limit, and states when access resumes. Raise or remove the limit to restore access sooner. + The rate limits for the Messages API are measured in requests per minute (RPM), input tokens per minute (ITPM), and output tokens per minute (OTPM) for each model class. If you exceed any of the rate limits, you will get a 429 error describing which rate limit was exceeded. When the response has a retry-after header, wait that many seconds before retrying. + Many API providers use a combined "tokens per minute" (TPM) limit that might include all tokens, both cached and uncached, input and output. For most Claude models, only uncached input tokens count toward your ITPM rate limits. This is a key advantage that makes the rate limits effectively higher than they might initially appear. + Example: With a 2,000,000 ITPM limit and an 80 percent cache hit rate, you could effectively process 10,000,000 total input tokens per minute (2M uncached + 8M cached), because cached tokens don't count toward your rate limit. + To get the highest throughput within your rate limits, cache repeated content such as system instructions and prompts, large context documents, tool definitions, and conversation history; see prompt caching for guidance. With effective caching, you can substantially increase your actual throughput without raising your rate limits. Monitor your cache hit rate on the Usage page in the Claude Console to tune your caching strategy. + The Message Batches API has its own set of rate limits, which are shared across all models. These include a requests per minute (RPM) limit to all API endpoints and a limit on the number of batch requests that can be in the processing queue at the same time. A "batch request" here refers to part of a Message Batch. You can create a Message Batch containing thousands of batch requests, each of which counts toward this limit. A batch request is considered part of the processing queue when it has yet to be successfully processed by the model. + Claude Managed Agents endpoints are rate-limited for each organization. These limits are separate from the Messages API rate limits described earlier on this page. + In addition to providing token and request charts, the Usage page provides two separate rate limit charts. Use these charts to see what headroom you have to grow, identify when you might be hitting peak use, understand what rate limits to request, and learn how to improve your caching rates. The charts visualize a number of metrics for a given rate limit (for example, for each model): + Setting lower limits for workspaces + To protect Workspaces in your Organization from potential overuse, you can set custom spend and rate limits for each Workspace. + Workspace limits are set for each limiter type (such as requests per minute, input tokens per minute, or output tokens per minute).
-
Tracking since Jun 25, 2026 · 124 snapshots on file
OpenAI API: 1 additive change
Full history & current snapshot →View raw diff +1 −0
+ New endpoint: DELETE /audio/voices/{voice_id} -
GitHub Copilot pricing changed
Tracking since Jun 25, 2026 · 10 snapshots on file
Free plan now explicitly includes 50 chat requests per month alongside 2,000 completions
Free · rate_limit2,000 completions per month→2,000 completions and 50 chat requests per monthFull history & current snapshot →View raw diff +12 −68
- GitHub Copilot transforms the developer experience. Backed by the leaders in AI, GitHub Copilot provides contextualized assistance throughout the software development lifecycle, from inline suggestions and chat assistance in the IDE to code explanations and answers to docs in GitHub and more. With GitHub Copilot elevating their workflow, developers can focus on: value, innovation, and happiness. - GitHub Copilot integrates with leading editors, including Visual Studio Code, Visual Studio, JetBrains IDEs, and Neovim, and, unlike other AI coding assistants, is natively built into GitHub. Growing to millions of individual users and tens of thousands of business customers, GitHub Copilot is the world’s most widely adopted AI developer tool and the competitive advantage developers ask for by name. - Who is eligible to access GitHub Copilot for free? - GitHub Copilot Free is a new free pricing tier with limited functionality for individual developers. Users assigned a Copilot Business or Copilot Enterprise seat are not eligible for access. Users with access to Copilot Pro through a paid subscription, trial, or through an existing verified OSS, student, faculty, or MVP account may elect to use Free instead. - GitHub Copilot is available as an extension in Visual Studio Code, Visual Studio, Vim, Neovim, the JetBrains suite of IDEs, and Azure Data Studio. Although inline suggestion functionality is available across all these extensions, chat functionality is currently available only in Visual Studio Code, JetBrains, and Visual Studio. GitHub Copilot is also supported in terminals through GitHub CLI and as a chat integration in Windows Terminal Canary. With the GitHub Copilot Enterprise plan, GitHub Copilot is natively integrated into GitHub.com. All plans are supported in GitHub Copilot in GitHub Mobile. GitHub Mobile for Copilot Pro and Copilot Business have access to Bing and public repository code search. Copilot Enterprise in GitHub Mobile gives you additional access to your organization's knowledge. - What are the differences between the GitHub Copilot Business, GitHub Copilot Enterprise, and GitHub Copilot Individual plans? - Organizations can choose between GitHub Copilot Business and GitHub Copilot Enterprise. GitHub Copilot Business primarily features GitHub Copilot in the coding environment - that is the IDE, CLI and GitHub Mobile. GitHub Copilot Enterprise includes everything in GitHub Copilot Business. It also adds an additional layer of customization for organizations and integrates into GitHub.com as a chat interface to allow developers to converse with GitHub Copilot throughout the platform. GitHub Copilot Enterprise can index an organization's codebase for a deeper understanding of the customer's knowledge for more tailored suggestions. For more details, refer to the documentation. - GitHub Copilot Individual is designed for individual developers, freelancers, students, educators, and open source maintainers. The plan includes all the features of GitHub Copilot Business except organizational license management, policy management, and IP indemnity. - GitHub Copilot is powered by generative AI models developed by GitHub, OpenAI, and Microsoft. It has been trained on natural language text and source code from publicly available sources, including code in public repositories on GitHub. Starting on April 24, GitHub may also use interactions from users with a Copilot Free, Copilot Pro, and Copilot Pro+ subscription - including inputs, outputs, code snippets, and associated context - to train and improve our AI models unless they have opted out. This allows us to build more intelligent, context-aware coding assistance for a more diverse set of coding tasks based on real-world development patterns. Users were notified 30 days before the change went into effect and can opt out from allowing their data to be used for training in their GitHub account settings at any time. - Which plan includes GitHub Copilot Autofix? - GitHub Copilot Autofix provides contextual explanations and code suggestions to help developers fix vulnerabilities in code, and is included in GitHub Advanced Security. - Access to Copilot Business and Enterprise is managed by your GitHub Administrator. They can control access to preview features, models, and set GitHub Copilot policies for your organization. Additionally, you can use your network firewall to explicitly allow access to Copilot Business and/or block access to Copilot Pro or Free. For more details, refer to the documentation. - Organizations can choose between GitHub Copilot Business and GitHub Copilot Enterprise. GitHub Copilot Business primarily features GitHub Copilot in the coding environment - that is the IDE, CLI and GitHub Mobile. GitHub Copilot Enterprise includes everything in GitHub Copilot Business. It also adds an additional layer of customization for organizations and integrates into GitHub.com as a chat interface to allow developers to converse with Copilot throughout the platform. GitHub Copilot Enterprise can index an organization’s codebase for a deeper understanding of the customer’s knowledge for more tailored suggestions and will offer customers access to fine-tuned custom, private models for code completion. - GitHub Copilot Pro is designed for individual developers, freelancers, students, educators, and open source maintainers. The plan includes all the features of GitHub Copilot Business except organizational license management, policy management, and IP indemnity. - GitHub Copilot Max is built for heavy Copilot usage, including sustained agent-driven workflows, and includes $100/month in GitHub AI Credits. - How can I upgrade my GitHub Copilot Free license to Copilot Pro? - If you're on the Free plan, you can upgrade to Pro through your Copilot settings page or directly on the Copilot marketing page. - GitHub Copilot Free users are limited to 2000 completions and 50 chat requests (including Copilot Edits). - Which plan includes GitHub Copilot Autofix? - GitHub Copilot Autofix provides contextual explanations and code suggestions to help developers fix vulnerabilities in code, and is included in GitHub Advanced Security and available to all public repositories. - Can users in my organization use Copilot code reviews for their pull requests if they don’t have a Copilot license? - Organizations can now enable Copilot code review on all pull requests on github.com—including pull requests from users who are not assigned a Copilot license. - This allows you to extend the quality and rich analysis of Copilot code review to all pull requests, regardless of its author, giving you complete coverage and confidence that pull requests have been reviewed. - To enable this functionality, an enterprise/org admin must first have Copilot enabled and then enabled two policies. - Usage from non-licensed users is billed directly to your organization as GitHub AI Credits. This flexible model allows you to get full review coverage on every PR without purchasing a full Copilot seat for non-development contributors who may not need Copilot. - No. This capability is off by default and gives the enterprise admin control to enable or disable. An admin must explicitly enable two separate policies to activate: - ‘GitHub AI Credits paid usage’ must be enabled to allow enterprises to be charged for GitHub AI Credits exceeding their included usage. - We encourage admins to set up budgets to control spending on our metered products, especially customers who have not enabled the ‘Premium request paid usage’ policy in the past. You can track all premium request usage in your billing dashboard to monitor and control spending. - You use credits when you chat with Copilot, work with agents, or use Copilot CLI, Spaces, and Spark. Code completions and next edit suggestions don't use credits. They remain unlimited with every paid plan. - Does GitHub use Copilot Business or Enterprise data to train GitHub’s model? - No. GitHub does not use either Copilot Business or Enterprise data to train its models. - How long does GitHub retain Copilot data for Business and Enterprise customers? - If and for how long GitHub’s retains Copilot data depends on how a Copilot user accesses Copilot and for what purpose. The default settings for Copilot Business and Enterprise Customers are as follows: - We are conducting internal testing of GitHub Copilot’s ease of use by developers with disabilities and working to ensure that GitHub Copilot is accessible to all developers. Please feel free to share your feedback on GitHub Copilot accessibility in our feedback forum. - GitHub Copilot transforms the developer experience. Backed by the leaders in AI, GitHub Copilot provides contextualized assistance throughout the software development lifecycle, from inline suggestions and chat assistance in the IDE to code explanations and answers to docs in GitHub and more. With GitHub Copilot elevating their workflow, developers can focus on: value, innovation, and happiness. - GitHub Copilot integrates with leading editors, including Visual Studio Code, Visual Studio, JetBrains IDEs, and Neovim, and, unlike other AI coding assistants, is natively built into GitHub. Growing to millions of individual users and tens of thousands of business customers, GitHub Copilot is the world’s most widely adopted AI developer tool and the competitive advantage developers ask for by name. - Who is eligible to access GitHub Copilot for free? - GitHub Copilot Free is a new free pricing tier with limited functionality for individual developers. Users assigned a Copilot Business or Copilot Enterprise seat are not eligible for access. Users with access to Copilot Pro through a paid subscription, trial, or through an existing verified OSS, student, faculty, or MVP account may elect to use Free instead. - GitHub Copilot is available as an extension in Visual Studio Code, Visual Studio, Vim, Neovim, the JetBrains suite of IDEs, and Azure Data Studio. Although inline suggestion functionality is available across all these extensions, chat functionality is currently available only in Visual Studio Code, JetBrains, and Visual Studio. GitHub Copilot is also supported in terminals through GitHub CLI and as a chat integration in Windows Terminal Canary. With the GitHub Copilot Enterprise plan, GitHub Copilot is natively integrated into GitHub.com. All plans are supported in GitHub Copilot in GitHub Mobile. GitHub Mobile for Copilot Pro and Copilot Business have access to Bing and public repository code search. Copilot Enterprise in GitHub Mobile gives you additional access to your organization's knowledge. - What are the differences between the GitHub Copilot Business, GitHub Copilot Enterprise, and GitHub Copilot Individual plans? - Organizations can choose between GitHub Copilot Business and GitHub Copilot Enterprise. GitHub Copilot Business primarily features GitHub Copilot in the coding environment - that is the IDE, CLI and GitHub Mobile. GitHub Copilot Enterprise includes everything in GitHub Copilot Business. It also adds an additional layer of customization for organizations and integrates into GitHub.com as a chat interface to allow developers to converse with GitHub Copilot throughout the platform. GitHub Copilot Enterprise can index an organization's codebase for a deeper understanding of the customer's knowledge for more tailored suggestions. For more details, refer to the documentation. - GitHub Copilot Individual is designed for individual developers, freelancers, students, educators, and open source maintainers. The plan includes all the features of GitHub Copilot Business except organizational license management, policy management, and IP indemnity. - GitHub Copilot is powered by generative AI models developed by GitHub, OpenAI, and Microsoft. It has been trained on natural language text and source code from publicly available sources, including code in public repositories on GitHub. Starting on April 24, GitHub may also use interactions from users with a Copilot Free, Copilot Pro, and Copilot Pro+ subscription - including inputs, outputs, code snippets, and associated context - to train and improve our AI models unless they have opted out. This allows us to build more intelligent, context-aware coding assistance for a more diverse set of coding tasks based on real-world development patterns. Users were notified 30 days before the change went into effect and can opt out from allowing their data to be used for training in their GitHub account settings at any time. - Which plan includes GitHub Copilot Autofix? - GitHub Copilot Autofix provides contextual explanations and code suggestions to help developers fix vulnerabilities in code, and is included in GitHub Advanced Security. - Access to Copilot Business and Enterprise is managed by your GitHub Administrator. They can control access to preview features, models, and set GitHub Copilot policies for your organization. Additionally, you can use your network firewall to explicitly allow access to Copilot Business and/or block access to Copilot Pro or Free. For more details, refer to the documentation. - Organizations can choose between GitHub Copilot Business and GitHub Copilot Enterprise. GitHub Copilot Business primarily features GitHub Copilot in the coding environment - that is the IDE, CLI and GitHub Mobile. GitHub Copilot Enterprise includes everything in GitHub Copilot Business. It also adds an additional layer of customization for organizations and integrates into GitHub.com as a chat interface to allow developers to converse with Copilot throughout the platform. GitHub Copilot Enterprise can index an organization’s codebase for a deeper understanding of the customer’s knowledge for more tailored suggestions and will offer customers access to fine-tuned custom, private models for code completion. - GitHub Copilot Pro is designed for individual developers, freelancers, students, educators, and open source maintainers. The plan includes all the features of GitHub Copilot Business except organizational license management, policy management, and IP indemnity. - GitHub Copilot Max is built for heavy Copilot usage, including sustained agent-driven workflows, and includes $100/month in GitHub AI Credits. - How can I upgrade my GitHub Copilot Free license to Copilot Pro? - If you're on the Free plan, you can upgrade to Pro through your Copilot settings page or directly on the Copilot marketing page. - GitHub Copilot Free users are limited to 2000 completions and 50 chat requests (including Copilot Edits). - Which plan includes GitHub Copilot Autofix? - GitHub Copilot Autofix provides contextual explanations and code suggestions to help developers fix vulnerabilities in code, and is included in GitHub Advanced Security and available to all public repositories. - Can users in my organization use Copilot code reviews for their pull requests if they don’t have a Copilot license? - Organizations can now enable Copilot code review on all pull requests on github.com—including pull requests from users who are not assigned a Copilot license. - This allows you to extend the quality and rich analysis of Copilot code review to all pull requests, regardless of its author, giving you complete coverage and confidence that pull requests have been reviewed. - To enable this functionality, an enterprise/org admin must first have Copilot enabled and then enabled two policies. - Usage from non-licensed users is billed directly to your organization as GitHub AI Credits. This flexible model allows you to get full review coverage on every PR without purchasing a full Copilot seat for non-development contributors who may not need Copilot. - No. This capability is off by default and gives the enterprise admin control to enable or disable. An admin must explicitly enable two separate policies to activate:
-
Anthropic model deprecations changed
Tracking since Jun 25, 2026 · 50 snapshots on file
Anthropic announced model retirements with July 21 2025 deadline for Claude 2.0, Claude 2.1, and Claude 3 Sonnet models, plus removed temperature/top_p/top_k parameters in API version 2026-10-21.
- July 21, 2025 retirement date for claude-2.0, claude-2.1, and claude-3-sonnet-20240229
- API version 2026-10-21 removes temperature, top_p, and top_k parameters returning 400 error
- November 6, 2024 retirement of Claude 1 and Instant models with migration to claude-haiku-4-5-20251001
Full history & current snapshot →View raw diff +18 −0
+ On January 21, 2025, Anthropic notified developers using Claude 2, Claude 2.1, and Claude Sonnet 3 models of their upcoming retirements. + | July 21, 2025 | claude-2.0 | claude-opus-4-8 | + | July 21, 2025 | claude-2.1 | claude-opus-4-8 | + | July 21, 2025 | claude-3-sonnet-20240229 | claude-sonnet-5-5 | + 2024-09-04: Claude 1 and Instant models + On September 4, 2024, Anthropic notified developers using Claude 1 and Instant models of their upcoming retirements. + | November 6, 2024 | claude-1.0 | claude-haiku-4-5-20251001 | + | November 6, 2024 | claude-1.1 | claude-haiku-4-5-20251001 | + | November 6, 2024 | claude-1.2 | claude-haiku-4-5-20251001 | + | November 6, 2024 | claude-1.3 | claude-haiku-4-5-20251001 | + | November 6, 2024 | claude-instant-1.0 | claude-haiku-4-5-20251001 | + | November 6, 2024 | claude-instant-1.1 | claude-haiku-4-5-20251001 | + | November 6, 2024 | claude-instant-1.2 | claude-haiku-4-5-20251001 | + API parameter deprecations + Anthropic occasionally deprecates request parameters that no longer apply to current models. A new model can reject a deprecated parameter, and a new API version can remove it for every model. + | Parameter | Status | Behavior | Recommended replacement | + | temperature, top_p, top_k | Removed in API version 2026-10-21 | Returns a 400 error. Earlier API versions return it when set to a non-default value on Claude 4.7 and later models and Claude Mythos Preview. | Omit and use prompting to guide model behavior. | + For migration steps, see the migration guide.
-
Tracking since Jun 25, 2026 · 30 snapshots on file
Arena Elo (text, overall) updated.
Full history & current snapshot →View raw diff +87 −87
- amazon-nova-experimental-chat-26-02-10 1448 - chatgpt-4o-latest-20250326 1428 - Claude Opus 4.6 1498 - Claude Opus 4.6 (High) 1502 - Claude Opus 5 (Max) 1508 - claude-opus-4-5-20251101 1451 - claude-opus-4-5-20251101-high-32k 1448 - claude-opus-4-8 1454 - claude-opus-4-8-high 1461 - claude-sonnet-4-5-20250929 1438 - claude-sonnet-4-5-20250929-high-32k 1435 - claude-sonnet-4-6 1458 - claude-sonnet-5-high 1442 - claude-sonnet-5.5-xhigh 1470 - deepseek-v4-flash 1432 - deepseek-v4-pro 1451 - deepseek-v4-pro-high-20260813 1446 - deepseek-v4-pro-high-preview 1445 - deepseek-v4.1-flash-max 1463 - dola-seed-2.0-pro 1448 - ernie-5.0-0110 1445 - ernie-5.0-preview-1022 1430 - ernie-5.0-preview-1203 1442 - ernie-5.1 1468 - Gemini 3.7 Flash (High) 1487 - gemini-2.5-pro 1458 - gemini-3-flash 1466 - gemini-3-flash (thinking-minimal) 1442 - gemini-3.5-flash-high 1482 - gemini-3.5-flash-lite 1435 - gemini-3.5-flash-medium 1477 - gemini-3.6-flash-high 1479 - gemma-4-26b-a4b 1434 - gemma-4-31b 1443 - glm-4.5 1430 - glm-4.6 1440 - glm-4.7 1435 - glm-5 1446 - glm-5.1 1462 - glm-5.2-max 1470 - glm-5.3-flash 1471 - glm-5.3-max 1471 - glm-5v-turbo 1438 - gpt-5.1-high 1443 - gpt-5.2-chat-latest-20260210 1438 - gpt-5.4 1452 - gpt-5.4-high 1469 - gpt-5.5 1467 - gpt-5.5-high 1471 - gpt-5.6-luna-xhigh 1431 - gpt-5.6-sol-xhigh 1457 - gpt-5.6-terra-xhigh 1447 - gpt-6-astra-max 1443 - gpt-6.1-sol-max 1448 - grok-4.1 1436 - grok-4.1-thinking 1436 - grok-4.20-beta-0309-reasoning 1451 - grok-4.20-beta1 1445 - grok-4.20-multi-agent-beta-0309 1450 - grok-4.5 1448
-
NVIDIA Nemotron models changed
Tracking since Jun 25, 2026 · 40 snapshots on file
Removed MiniMax-M3-DSpark model with 'other' license (expiring 2026-07-15) from NVIDIA Nemotron offerings.
- nvidia/MiniMax-M3-DSpark
- license:other
- 2026-07-15
Full history & current snapshot →View raw diff +0 −1
- nvidia/MiniMax-M3-DSpark license:other 2026-07-15 -
Tracking since Jun 25, 2026 · 20 snapshots on file
Microsoft Phi model changed from Dayhoff-170M-GR-16000 to Sentiment-Analysis-7B with license expiration date moving from 2026-01-24 to 2026-10-17.
- Model name changed from microsoft/Dayhoff-170M-GR-16000 to microsoft/Sentiment-Analysis-7B
- License expiration date changed from 2026-01-24 to 2026-10-17
Full history & current snapshot →View raw diff +1 −1
- microsoft/Dayhoff-170M-GR-16000 license:mit 2026-01-24 + microsoft/Sentiment-Analysis-7B license:mit 2026-10-17
-
NVIDIA Nemotron models changed
Tracking since Jun 25, 2026 · 40 snapshots on file
Removed two NVIDIA models and added Flora-Captioner model with unknown license status, with effective date of 2026-10-01.
- Removed nvidia/Ising-Calibration-1.5-31B-BF16
- Removed nvidia/Kimi-K2.7-Code-DFlash
- Added nvidia/Flora-Captioner with license:? dated 2026-10-01
Full history & current snapshot →View raw diff +1 −2
- nvidia/Ising-Calibration-1.5-31B-BF16 license:other 2026-07-13 - nvidia/Kimi-K2.7-Code-DFlash license:other 2026-07-08 + nvidia/Flora-Captioner license:? 2026-10-01
-
Tracking since Jun 25, 2026 · 66 snapshots on file
Clarified that workspace spend limits (not rate limits) cannot be set on the default workspace.
Full history & current snapshot →View raw diff +1 −1
- You can't set limits on the default Workspace. + You can't set spend limits on the default Workspace.
-
Mistral models & deprecations changed
Tracking since Jun 25, 2026 · 18 snapshots on file
Four Voxtral v3.0 models added: Mini Transcribe 3, Small Transcribe 3, Mini Transcribe Realtime 3, and TTS 3 with updated capabilities.
- Voxtral Mini Transcribe 3 for competitive transcription and diarization
- Voxtral Small Transcribe 3 as flagship model with SOTA accuracy
- Voxtral Mini Transcribe Realtime 3 for agentic use-cases with configurable latency
Full history & current snapshot →View raw diff +12 −0
+ Voxtral Mini Transcribe 3 + Strong transcription and diarization at a competitive price point. + v3.0 + Voxtral Small Transcribe 3 + Flagship transcription and diarization model with SOTA accuracy. + v3.0 + Voxtral Mini Transcribe Realtime 3 + Realtime transcription model for agentic use-cases with configurable latency and turn detection. + v3.0 + Voxtral TTS 3 + State-of-the-art TTS with zero-shot and professional voice cloning, 9 languages, and streaming. + v3.0
-
Tracking since Jun 25, 2026 · 18 snapshots on file
Clarification that deepseek-flash model is now served by DeepSeek-V4.1-Flash-1017 with unchanged calling method.
- deepseek-flash is now served by DeepSeek-V4.1-Flash-1017
- legacy names deepseek-v4-flash and deepseek-v4-flash-vision-exp still accepted but retired models
- requests billed at the Flash price
Full history & current snapshot →View raw diff +2 −30
- Model prices are listed in units of per 1M tokens. A token, the smallest unit of text that the model recognizes, can be a word, a number, or even a punctuation mark. We bill model usage based on the total number of input and output tokens by the model. - DeepSeek-V4-Pro-0813 - 1M INPUT TOKENS - $0.003 - $0.022 - $0.006 - $0.044 - 1M INPUT TOKENS - $0.15 - $0.66 - $0.3 - $1.32 - 1M OUTPUT TOKENS - $0.6 - $1.98 - $1.2 - $3.96 - Concurrency Limit(3) - (1) Use deepseek-flash as the model name. The legacy names deepseek-v4-flash and deepseek-v4-flash-vision-exp are still accepted, but the corresponding models have been retired, their requests are served by the DeepSeek-V4.1-Flash-1017 model and billed at the Flash price. - (3) For more details on concurrency limits, please refer to Rate Limit & Isolation. - Search and Fetch are standalone APIs billed per request. - Price - Per 1,000 successful requests - $0.75 - Per 1,000 successful requests - $0.75 - Note: when you use the DeepSeek API through DeepSeek Harness, Claude Code, or Codex, server-side tool calls already hit the Search and Fetch APIs. These calls are currently free of charge; per-request billing starts at 00:00 Beijing time on November 1, 2026. - For model usage, the expense = number of tokens × price. Search and Fetch are charged per successful request at the rates above. - Product prices may vary and DeepSeek reserves the right to adjust them. We recommend topping up based on your actual usage and regularly checking this page for the most recent pricing information. - Token & Token Usage + (1) Use deepseek-flash as the model name. It is now served by DeepSeek-V4.1-Flash-1017, and the calling method remains unchanged. The legacy names deepseek-v4-flash and deepseek-v4-flash-vision-exp are still accepted, but the corresponding models have been retired; their requests are served by DeepSeek-V4.1-Flash-1017 and billed at the Flash price. + -H "Authorization: Bearer ${DEEPSEEK_API_KEY}" \
-
Tracking since Jun 25, 2026 · 20 snapshots on file
Removed microsoft/Dayhoff-3b-GR-HM-41000 model with MIT license from available models.
- microsoft/Dayhoff-3b-GR-HM-41000 model removed
- MIT license
- Dated 2026-01-24
Full history & current snapshot →View raw diff +0 −1
- microsoft/Dayhoff-3b-GR-HM-41000 license:mit 2026-01-24 -
Google Gemini deprecations changed
Tracking since Jun 25, 2026 · 25 snapshots on file
Model deprecation and migration paths updated: Gemini flash-lite models now migrate to gemini-3.1-flash-lite, Imagen and Veo models redirect to gemini-nano-banana-2.1 or gemini-omni-1.1-flash, and new robotics models added with no shutdown date announced.
- gemini-2.0-flash-lite-preview migration target changed from gemini-2.5-flash-lite to gemini-3.1-flash-lite
- veo-3.0-generate-001 and veo-3.0-fast-generate-001 now redirect to gemini-omni-1.1-flash instead of veo-3.1-generate-preview
- gemini-robotics-er-2-preview and gemini-robotics-er-2-streaming-preview added with July 30, 2026 release date and no shutdown date announced
Full history & current snapshot →View raw diff +14 −12
- | gemini-2.0-flash-lite-preview | February 5, 2025 | December 9, 2025 | gemini-2.5-flash-lite | - | gemini-2.0-flash-lite-preview-02-05 | February 5, 2025 | December 9, 2025 | gemini-2.5-flash-lite | - | imagen-3.0-generate-002 | February 6, 2025 | November 10, 2025 | imagen-4.0-generate-001 | - | imagen-4.0-generate-preview-06-06 | June 24, 2025 | February 17, 2026 | imagen-4.0-generate-001 | - | imagen-4.0-ultra-generate-preview-06-06 | June 24, 2025 | February 17, 2026 | imagen-4.0-ultra-generate-001 | - | veo-3.0-generate-001 | September 9, 2025 | June 30, 2026 | veo-3.1-generate-preview or the GA models on the Gemini Enterprise Agent Platform | - | veo-3.0-fast-generate-001 | September 9, 2025 | June 30, 2026 | veo-3.1-fast-generate-preview or the GA models on the Gemini Enterprise Agent Platform | - | veo-2.0-generate-001 | April 9, 2025 | June 30, 2026 | veo-3.1-generate-preview or the GA models on the Gemini Enterprise Agent Platform | - | veo-3.0-generate-preview | July 31, 2025 | November 12, 2025 | veo-3.1-generate-preview | - | veo-3.0-fast-generate-preview | July 31, 2025 | November 12, 2025 | veo-3.1-fast-generate-preview | - | gemini-robotics-er-1.6-preview | April 14, 2026 | August 31, 2026 | gemini-robotics-er-2-preview | - | gemini-robotics-er-1.5-preview | September 25, 2025 | April 30, 2026 | gemini-robotics-er-1.6-preview | + | gemini-2.0-flash-lite-preview | February 5, 2025 | December 9, 2025 | gemini-3.1-flash-lite | + | gemini-2.0-flash-lite-preview-02-05 | February 5, 2025 | December 9, 2025 | gemini-3.1-flash-lite | + | imagen-3.0-generate-002 | February 6, 2025 | November 10, 2025 | gemini-nano-banana-2.1 | + | imagen-4.0-generate-preview-06-06 | June 24, 2025 | February 17, 2026 | gemini-nano-banana-2.1 | + | imagen-4.0-ultra-generate-preview-06-06 | June 24, 2025 | February 17, 2026 | gemini-nano-banana-2.1 | + | veo-3.0-generate-001 | September 9, 2025 | June 30, 2026 | gemini-omni-1.1-flash or the GA models on the Gemini Enterprise Agent Platform | + | veo-3.0-fast-generate-001 | September 9, 2025 | June 30, 2026 | gemini-omni-1.1-flash or the GA models on the Gemini Enterprise Agent Platform | + | veo-2.0-generate-001 | April 9, 2025 | June 30, 2026 | gemini-omni-1.1-flash or the GA models on the Gemini Enterprise Agent Platform | + | veo-3.0-generate-preview | July 31, 2025 | November 12, 2025 | gemini-omni-1.1-flash | + | veo-3.0-fast-generate-preview | July 31, 2025 | November 12, 2025 | gemini-omni-1.1-flash | + | gemini-robotics-er-2-preview | July 30, 2026 | No shutdown date announced | | + | gemini-robotics-er-2-streaming-preview | July 30, 2026 | No shutdown date announced | | + | gemini-robotics-er-1.6-preview | April 14, 2026 | August 31, 2026 | gemini-robotics-er-2-preview | + | gemini-robotics-er-1.5-preview | September 25, 2025 | April 30, 2026 | gemini-robotics-er-2-preview |
-
Meta reported benchmarks updated
Tracking since Jun 25, 2026 · 17 snapshots on file
Llama 4 Maverick: 12 benchmark claims (via web search)
Full history & current snapshot → -
Google Gemini API pricing changed
Tracking since Jun 25, 2026 · 30 snapshots on file
Multiple Gemini models seeing 2x price increases in late 2026/early 2027; Veo 3.1 Preview deprecated
Gemini 4 Argon pricing increases December 1, 2026: input $2→$4/1M, output $10→$20/1M tokensGemini 2.0 Flash pricing increases January 1, 2027: input $0.75→$1.50/1M, output $3.75→$7.50/1M tokensVeo 3.1 Preview models deprecated, shutting down October 22, 2026Full history & current snapshot →View raw diff +1 −0
+ Warning: Veo 3.1 Preview (veo-3.1-generate-preview, veo-3.1-fast-generate-preview, and veo-3.1-lite-generate-preview) is deprecated and will be shut down on October 22, 2026; migrate to Gemini Omni Flash (gemini-omni-1.1-flash) or the GA models available through the Gemini Enterprise Agent Platform to avoid service disruption. -
GitHub Copilot models deprecated on October 19, 2026
As of today, October 19, 2026, we have deprecated the following models across most GitHub Copilot experiences (including Copilot Chat, inline edits, ask and agent modes, and code completions). These… The post GitHub Copilot models deprecated on October 19, 2026 appeared first on The GitHub Blog.
Full history & current snapshot → -
Tracking since Jun 25, 2026 · 6 snapshots on file
Pricing tier targeting expanded from Small Business/Enterprise to include Entrepreneurs, Agencies, and All Solutions categories; promotional messaging for GLM 5.3 Flash removed.
- Added tier categories: Entrepreneurs, Agencies, All Solutions
- Removed promotional message: GLM 5.3 Flash free this weekend
Full history & current snapshot →View raw diff +1 −2
- Small BusinessEnterprise - 🎉GLM 5.3 Flash: free this weekend + EntrepreneursSmall BusinessAgenciesEnterpriseAll Solutions
-
Tracking since Jun 25, 2026 · 25 snapshots on file
DigitalOcean API: 3 additive changes
Full history & current snapshot →View raw diff +3 −0
+ New endpoint: GET /v2/insights/resources + New endpoint: GET /v2/insights/resources/{resource_urn} + New endpoint: PUT /v2/insights/resources/{resource_urn}/signals
-
Moonshot AI reported benchmarks updated
Tracking since Jun 25, 2026 · 17 snapshots on file
Kimi K3: 12 benchmark claims (via web search)
Full history & current snapshot → -
Zhipu AI (Z.ai) reported benchmarks updated
Tracking since Jun 25, 2026 · 17 snapshots on file
GLM-5.3: 6 benchmark claims (via web search)
Full history & current snapshot → -
Alibaba reported benchmarks updated
Tracking since Jun 25, 2026 · 17 snapshots on file
Qwen3.8-Max: 11 benchmark claims (via web search)
Full history & current snapshot → -
DeepSeek reported benchmarks updated
Tracking since Jun 25, 2026 · 17 snapshots on file
DeepSeek-V3.2: 10 benchmark claims (via web search)
Full history & current snapshot → -
xAI reported benchmarks updated
Tracking since Jun 25, 2026 · 17 snapshots on file
Grok 4.7: 7 benchmark claims (via web search)
Full history & current snapshot → -
Google reported benchmarks updated
Tracking since Jun 25, 2026 · 17 snapshots on file
Gemini 3.1 Pro: 10 benchmark claims (via web search)
Full history & current snapshot → -
OpenAI reported benchmarks updated
Tracking since Jun 25, 2026 · 17 snapshots on file
GPT-6 Astra: 10 benchmark claims (via web search)
Full history & current snapshot → -
Anthropic reported benchmarks updated
Tracking since Jun 25, 2026 · 17 snapshots on file
Claude Opus 5.5: 12 benchmark claims (via web search)
Full history & current snapshot → -
NVIDIA Nemotron models changed
Tracking since Jun 25, 2026 · 40 snapshots on file
Removed 6 NVIDIA Nemotron and SDLLM model entries from the catalog.
- nvidia/Nemotron-Labs-Audex-2B removed
- nvidia/Nemotron-Labs-Audex-30B-A3B removed
- 4 SDLLM variants removed (AR, Duo, EsoLM, MDLM)
Full history & current snapshot →View raw diff +0 −6
- nvidia/Nemotron-Labs-Audex-2B license:other 2026-07-06 - nvidia/Nemotron-Labs-Audex-30B-A3B license:other 2026-07-06 - nvidia/SDLLM-AR-1.7B-Base license:cc-by-nc-4.0 2026-07-02 - nvidia/SDLLM-Duo-1.7B-Base license:cc-by-nc-4.0 2026-07-02 - nvidia/SDLLM-EsoLM-1.7B-Base license:cc-by-nc-4.0 2026-07-02 - nvidia/SDLLM-MDLM-1.7B-Base license:cc-by-nc-4.0 2026-07-02
-
Windsurf (now Devin) pricing changed
Tracking since Jun 25, 2026 · 7 snapshots on file
SWE-2 free access expiration dates removed from Free and Pro plans
Free · notesFree use of SWE-2 Free in Devin Desktop and CLI through October 16, 2026→Free use of SWE-2Pro · notesFree SWE-2Free in Devin Desktop and CLI through October 16, 2026→Free SWE-2Full history & current snapshot →View raw diff +2 −2
- Free use of SWE-2 Free in Devin Desktop and CLI through October 16, 2026 - Free SWE-2Free in Devin Desktop and CLI through October 16, 2026 + Free use of SWE-2 + Free SWE-2
-
NVIDIA Nemotron models changed
Tracking since Jun 25, 2026 · 40 snapshots on file
Added Nemotron-3.5-Parse model with openmdw-1.1 license and expiration date of 2026-09-03.
- nvidia/Nemotron-3.5-Parse
- license:openmdw-1.1
- expiration: 2026-09-03
Full history & current snapshot →View raw diff +1 −0
+ nvidia/Nemotron-3.5-Parse license:openmdw-1.1 2026-09-03 -
Tracking since Jun 25, 2026 · 25 snapshots on file
Removed a model deprecation deadline row for gpt-4-1106-preview dated 2026-03-26.
- gpt-4-1106-preview deprecation deadline removed
- 2026-03-26 date removed
Full history & current snapshot →View raw diff +0 −1
- | 2026‑03‑26 | gpt-4-1106-preview | gpt-5 or gpt-4.1* | -
DeepSeek (weights) models changed
Tracking since Jun 25, 2026 · 9 snapshots on file
DeepSeek model updated from deepseek-coder-5.7bmqa-base to DeepSeek-V4.1-Flash-1017 with license change from other to mit and release date updated to 2026-10-17.
- Model name changed to DeepSeek-V4.1-Flash-1017
- License changed from other to mit
- Release date updated from 2023-10-31 to 2026-10-17
Full history & current snapshot →View raw diff +1 −1
- deepseek-ai/deepseek-coder-5.7bmqa-base license:other 2023-10-31 + deepseek-ai/DeepSeek-V4.1-Flash-1017 license:mit 2026-10-17
-
Tracking since Jun 25, 2026 · 6 snapshots on file
GLM 5.3 Flash model is free this weekend.
- GLM 5.3 Flash
- free this weekend
Full history & current snapshot →View raw diff +1 −0
+ 🎉GLM 5.3 Flash: free this weekend -
Tracking since Jun 25, 2026 · 18 snapshots on file
Legacy deepseek-v4-flash model names now served by DeepSeek-V4.1-Flash-1017 instead of DeepSeek-V4.1-Flash while maintaining Flash price billing.
DeepSeek-V4.1-Flash model name updated to DeepSeek-V4.1-Flash-1017 (legacy name change note only, no price change)Full history & current snapshot →View raw diff +1 −1
- (1) Use deepseek-flash as the model name. The legacy names deepseek-v4-flash and deepseek-v4-flash-vision-exp are still accepted, but the corresponding models have been retired, their requests are served by the DeepSeek-V4.1-Flash model and billed at the Flash price. + (1) Use deepseek-flash as the model name. The legacy names deepseek-v4-flash and deepseek-v4-flash-vision-exp are still accepted, but the corresponding models have been retired, their requests are served by the DeepSeek-V4.1-Flash-1017 model and billed at the Flash price.
-
Tracking since Jun 25, 2026 · 313 snapshots on file
ChatGPT Business and Enterprise plans added to pricing page with Standard ($20/mo) and Premium ($100/mo) seat options
ChatGPT Business and Enterprise plans now featured on business pricing page (previously API-only)Full history & current snapshot →View raw diff +214 −43
- Business Pricing | OpenAI - Price - Input:$10.00 / 1M tokens Cached input:$1.00 / 1M tokens Output:$50.00 / 1M tokens - Price - Input:$2.00 / 1M tokens Cached input:$0.10 / 1M tokens Output:$10.00 / 1M tokens - Price - Input:$0.10 / 1M tokens Cached input:$0.01 / 1M tokens Output:$0.50 / 1M tokens - Price - Image:$8.00 / 1M tokens for inputs$2.00 / 1M tokens for cached inputs$30.00 / 1M tokens for outputs Text:$5.00 / 1M tokens for inputs$1.25 / 1M tokens for cached inputs - Price - Image:$8.00 / 1M tokens for inputs$2.00 / 1M tokens for cached inputs$30.00 / 1M tokens for outputs Text:$5.00 / 1M tokens for inputs$1.25 / 1M tokens for cached inputs - Price - $0.05 per minute / $0.00083 per second - Price - Audio:$32.00 / 1M tokens for inputs$0.40 / 1M tokens for cached inputs$64.00 / 1M tokens for outputs Text:$4.00 / 1M tokens for inputs$0.40 / 1M tokens for cached inputs$24.00 / 1M tokens for outputs Image:$5.00 / 1M tokens for inputs$0.50 / 1M tokens for cached inputs - Price - Audio:$10.00 / 1M tokens for inputs$0.30 / 1M tokens for cached inputs$20.00 / 1M tokens for outputs Text:$0.60 / 1M tokens for inputs$0.06 / 1M tokens for cached inputs$2.40 / 1M tokens for outputs Image:$0.80 / 1M tokens for inputs$0.08 / 1M tokens for cached inputs - Price - $0.017 per minute / $0.00028 per second - Price - $0.0045 per minute / $0.00008 per second - Price - $0.034 per minute / $0.00057 per second - Price - $10.00 / 1k calls Search content tokens are free. - Price - Now:1 GB for $0.03 / 64GB for $1.92 per container Starting March 31, 2026:1 GB for $0.03 / 64GB for $1.92 per 20-minute session per container - Provides lower costs for requests in exchange for slower response times and occasional resource unavailability. Ideal for non-production or lower priority tasks. - Enterprise offerings - Contact our sales team to learn more about Data residency(opens in a new window), Scale Tier andReserved Capacity designed for cutting-edge customers running larger workloads. - We recommend experimenting with all of these models in the Playground(opens in a new window) to explore which models provide the best price performance trade-off for your usage. - Do you offer an enterprise package or SLAs? - We offer different tiers of access to our enterprise customers that include SLAs, lower latency, and more. Please contact our sales team to learn more. - Yes, we treat Playground usage the same as regular API usage. You will be billed at the per-token input and output prices mentioned above. - How will I know how many tokens I’ve used each month? - A token is a mathematical representation of natural language. Log in to your account to view your usage tracking dashboard(opens in a new window). This dashboard will show you how many tokens you’ve used during the current and past billing cycles. - You can set a monthly budget in your billing settings(opens in a new window), after which we’ll stop serving your requests. There may be a delay in enforcing the limit, and you are responsible for any overage incurred. You can also configure an email notification threshold to receive an email alert once you cross that threshold each month. We recommend checking your usage tracking dashboard(opens in a new window) regularly to monitor your spend. - Is access to the API included in ChatGPT Plus, Business, Enterprise or Edu? - No, OpenAI APIs are billed separately from ChatGPT Plus, Business, Enterprise and Edu. ChatGPT subscription pricing can be found at openai.com/chatgpt/pricing/. - Images are converted into tokens and charged per token. Text models price image tokens at standard text token rates, while GPT Image and gpt-realtime uses a separate image token rate. Models like gpt-4.1-mini, gpt-4.1-nano, and o4-mini convert images into tokens differently. Learn more in our docs(opens in a new window). - ChatGPT Business(opens in a new window) - ChatGPT Enterprise(opens in a new window) - ChatGPT for Education(opens in a new window) + ChatGPT Business(opens in a new window) + ChatGPT Enterprise(opens in a new window) + ChatGPT for Education(opens in a new window) + Image 1 + Business + A secure workspace with company context and flexible seat types for any budget. + Get started(opens in a new window) + Standard seat + $20/ month + $20/month if billed annually. $25/month if billed monthly. + Premium seat + $100/ month + 5x more usage than standard, with no 5-hour limit + $100/month if billed annually. $125/month if billed monthly. + No training on your business data by default + Mix and match seat types + For teams of 2–200 employees. Unlimited subject to abuse guardrails. Learn more(opens in a new window)
-
Tracking since Jun 25, 2026 · 18 snapshots on file
Search and Fetch API billing start date confirmed for November 1, 2026
Search and Fetch API billing clarified: free through Oct 31, 2026; paid billing starts Nov 1, 2026 (Beijing time)Full history & current snapshot →View raw diff +1 −1
- Web search in Claude Code and Codex: Search requests made through Claude Code's built-in Web Search tool or Codex's web_search feature will not incur Search API request fees during the first two weeks after the Search API launch. Standard Search API pricing applies after that period; the billing start date will be announced separately. + Note: when you use the DeepSeek API through DeepSeek Harness, Claude Code, or Codex, server-side tool calls already hit the Search and Fetch APIs. These calls are currently free of charge; per-request billing starts at 00:00 Beijing time on November 1, 2026.
-
NVIDIA Nemotron models changed
Tracking since Jun 25, 2026 · 40 snapshots on file
New model variant added: nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-BF16 with other license and availability date 2026-10-16.
- Model: nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-BF16
- License type: other
- Availability date: 2026-10-16
Full history & current snapshot →View raw diff +1 −0
+ nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-BF16 license:other 2026-10-16 -
Tracking since Jun 25, 2026 · 27 snapshots on file
Cursor added Cloud Agent Canvases, Environments API beta support, and increased cloud environment capacity to 300 repositories with 5x faster load times.
- Cloud agents can now build interactive canvases and share via link
- Environments API available in beta for programmatic management of cloud environments
- Cloud agent environments support up to 300 repositories and load 5x faster
Full history & current snapshot →View raw diff +18 −24
- Sep 2, 2026 · Changelog - Self-hosted machines - Cursor supports self-hosted machines, which let you keep tool execution entirely in your own network. - Your codebase, build outputs, and secrets all stay on internal machines running in your infrastructure, while the agent handles tool calls locally. - #Dynamic pool scheduling - My Machines connects a single laptop or VM to your account for personal workflows. - Team pools are named queues of workers for a team or enterprise. Capacity can grow as requests arrive and shrink when workers disconnect, so your self-hosted machines can scale with demand. Pools are not tied to one repository: name the pool, and any available worker can claim the request. - Pools can also hibernate idle machines, then restore within a reconnect window when a follow-up arrives, so you don't keep expensive capacity warm just for the next prompt. - #Run on your sandboxes - Cloud agents can now execute on infrastructure you already use, including from AWS Lambda, Coder, Cloudflare, Daytona, Modal, Namespace, Vercel, and E2B. - #Computer use on Linux and Mac - Self-hosted workers now support computer use on Linux and Mac. With the right desktop packages, an agent can click, type, take screenshots, and drive the browser. You can watch its desktop or take control from Cursor. - Sep 2, 2026 · Changelog - Self-hosted machines - Cursor supports self-hosted machines, which let you keep tool execution entirely in your own network. - Your codebase, build outputs, and secrets all stay on internal machines running in your infrastructure, while the agent handles tool calls locally. - #Dynamic pool scheduling - My Machines connects a single laptop or VM to your account for personal workflows. - Team pools are named queues of workers for a team or enterprise. Capacity can grow as requests arrive and shrink when workers disconnect, so your self-hosted machines can scale with demand. Pools are not tied to one repository: name the pool, and any available worker can claim the request. - Pools can also hibernate idle machines, then restore within a reconnect window when a follow-up arrives, so you don't keep expensive capacity warm just for the next prompt. - #Run on your sandboxes - Cloud agents can now execute on infrastructure you already use, including from AWS Lambda, Coder, Cloudflare, Daytona, Modal, Namespace, Vercel, and E2B. - #Computer use on Linux and Mac - Self-hosted workers now support computer use on Linux and Mac. With the right desktop packages, an agent can click, type, take screenshots, and drive the browser. You can watch its desktop or take control from Cursor. + Oct 16, 2026 · Changelog + Cloud Agent Canvases and Environment Improvements + #Canvases from cloud agents + Cloud agents can now build interactive canvases as part of their work. Share a finished canvas with your team by sending its link. + Learn more about sharing canvases. + #Environments API + The Environments API is now available in beta. Use it to programmatically create and manage the cloud environments your agents run in. + #Larger, faster environments + Cloud agent environments now support up to 300 repositories. Large environments also load 5x faster. + Oct 16, 2026 · Changelog + Cloud Agent Canvases and Environment Improvements + #Canvases from cloud agents + Cloud agents can now build interactive canvases as part of their work. Share a finished canvas with your team by sending its link. + Learn more about sharing canvases. + #Environments API + The Environments API is now available in beta. Use it to programmatically create and manage the cloud environments your agents run in. + #Larger, faster environments + Cloud agent environments now support up to 300 repositories. Large environments also load 5x faster.
-
Google Gemini API pricing changed
Tracking since Jun 25, 2026 · 30 snapshots on file
Clarified Gemini 4 Argon is unavailable on Free tier only (Google AI Pro restriction removed from note)
Gemini 4 Argon availability note updated: removed mention of Google AI Pro plan restrictionFull history & current snapshot →View raw diff +1 −1
- Note: Gemini 4 Argon is not available on the Free tier or on the Google AI Pro plan. + Note: Gemini 4 Argon is not available on the Free tier.
-
Tracking since Jun 25, 2026 · 5 snapshots on file
Fixed code examples formatting in API documentation; no changes to concurrency limits themselves.
Full history & current snapshot →View raw diff +4 −4
- "messages": {"role": "user", "content": "Hello!"}, - extra_body={"user_id": "your_user_id"} - "messages": {"role": "user", "content": "Hello!"}, - max_tokens=1024 + "messages": [{"role": "user", "content": "Hello!"}], + extra_body={"user_id": "your_user_id"}, + "messages": [{"role": "user", "content": "Hello!"}], + max_tokens=1024,
-
Tracking since Jun 25, 2026 · 124 snapshots on file
OpenAI API: 3 breaking changes
Full history & current snapshot →View raw diff +0 −3
- New required body field: type - New required body field: name - New required body field: prompt
-
Tracking since Jun 25, 2026 · 18 snapshots on file
DeepSeek API pricing page now displays model prices in per-1M-token units with specific rates for DeepSeek-V4-Pro-0813 and other models, plus Search and Fetch API pricing at $0.75 per 1,000 requests.
now$0.00$0.02$0.04$0.15$0.66$0.3$1.32$0.6Full history & current snapshot →View raw diff +29 −1
- -H "Authorization: Bearer ${DEEPSEEK_API_KEY}" \ + Model prices are listed in units of per 1M tokens. A token, the smallest unit of text that the model recognizes, can be a word, a number, or even a punctuation mark. We bill model usage based on the total number of input and output tokens by the model. + DeepSeek-V4-Pro-0813 + 1M INPUT TOKENS + $0.003 + $0.022 + $0.006 + $0.044 + 1M INPUT TOKENS + $0.15 + $0.66 + $0.3 + $1.32 + 1M OUTPUT TOKENS + $0.6 + $1.98 + $1.2 + $3.96 + Concurrency Limit(3) + (3) For more details on concurrency limits, please refer to Rate Limit & Isolation. + Search and Fetch are standalone APIs billed per request. + Price + Per 1,000 successful requests + $0.75 + Per 1,000 successful requests + $0.75 + Web search in Claude Code and Codex: Search requests made through Claude Code's built-in Web Search tool or Codex's web_search feature will not incur Search API request fees during the first two weeks after the Search API launch. Standard Search API pricing applies after that period; the billing start date will be announced separately. + For model usage, the expense = number of tokens × price. Search and Fetch are charged per successful request at the rates above. + Product prices may vary and DeepSeek reserves the right to adjust them. We recommend topping up based on your actual usage and regularly checking this page for the most recent pricing information. + Token & Token Usage
-
NVIDIA Nemotron models changed
Tracking since Jun 25, 2026 · 40 snapshots on file
NVIDIA Nemotron Parse 2.0 model listing was removed.
- nvidia/NVIDIA-Nemotron-Parse-2.0 removed
- license openmdw-1.1 no longer present
- expiration date 2026-06-30 no longer applicable
Full history & current snapshot →View raw diff +0 −1
- nvidia/NVIDIA-Nemotron-Parse-2.0 license:openmdw-1.1 2026-06-30 -
Tracking since Jun 25, 2026 · 25 snapshots on file
DigitalOcean API: 10 additive changes
Full history & current snapshot →View raw diff +10 −0
+ New endpoint: GET /tools + New endpoint: POST /tools/search + New endpoint: POST /tools/invoke + New endpoint: POST /code/execute + New endpoint: POST /approvals/session/{session_id}/{approval_id} + New endpoint: POST /approvals/{approval_id} + New endpoint: POST /v1/audio/speech + New endpoint: POST /v1/videos + New endpoint: GET /v1/videos/{video_id} + New endpoint: GET /v1/videos/{video_id}/content
-
GPQA Diamond (Epoch) scores changed
Tracking since Jun 25, 2026 · 35 snapshots on file
GPQA Diamond accuracy (0–1) updated.
Full history & current snapshot →View raw diff +1 −0
+ Muse-Spark-Open-816B-A42B_xhigh 0.886 -
Tracking since Jun 25, 2026 · 30 snapshots on file
Arena Elo (text, overall) updated.
Full history & current snapshot →View raw diff +31 −31
- claude-fable-5-high 1491 - claude-fable-5.1-max 1510 - claude-opus-4-6 1498 - claude-opus-4-6-high 1502 - claude-opus-4-7 1483 - claude-opus-4-7-high 1490 - claude-opus-5-high 1502 - claude-opus-5-max 1507 - claude-opus-5.5-high 1516 - claude-sonnet-5.5-xhigh 1471 - deepseek-v4-pro-high-20260813 1445 - gemini-3-pro 1479 - gemini-3.1-pro-preview 1480 - gemini-3.5-flash-medium 1478 - gemini-3.7-flash-high 1487 - gemini-3.8-flash-high 1499 - gemini-4-argon-high 1533 - glm-5.1 1463 - gpt-6-astra-max 1442 - gpt-6.1-sol-max 1447 - hy3 1437 - kimi-k3-max 1476 - mimo-v2.6-pro 1490 - muse-spark 1474 - muse-spark-1.1 1478 - muse-spark-1.2 (xHigh) 1485 - muse-spark-1.3-max 1490 - muse-wonder-max 1504 - nvidia-nemotron-3-ultra-550b-a55b-nvfp4 1444 - qwen3.8-max 1482 - Step 5 Preview 1440 + Claude Fable 5 (High) 1491 + Claude Fable 5.1 (Max) 1512 + Claude Opus 4.6 1498 + Claude Opus 4.6 (High) 1502 + Claude Opus 4.7 1483 + Claude Opus 4.7 (High) 1490 + Claude Opus 5 (High) 1503 + Claude Opus 5 (Max) 1508 + Claude Opus 5.5 (High) 1516 + claude-sonnet-5.5-xhigh 1470 + deepseek-v4-pro-high-20260813 1446 + Gemini 3 Pro 1479 + Gemini 3.1 Pro Preview 1480 + Gemini 3.7 Flash (High) 1487 + Gemini 3.8 Flash (High) 1499 + Gemini 4 Argon (High) 1532 + gemini-3.5-flash-medium 1477 + glm-5.1 1462 + gpt-6-astra-max 1443 + gpt-6.1-sol-max 1448 + hy3 1438 + Kimi K3 (Max) 1475 + mimo-v2.6-pro 1491 + Muse Spark 1474 + Muse Spark 1.1 1478 + Muse Spark 1.2 (xHigh) 1484 + Muse Spark 1.3 (Max) 1490 + Muse Wonder (Max) 1501 + nvidia-nemotron-3-ultra-550b-a55b-nvfp4 1445
-
NVIDIA Nemotron models changed
Tracking since Jun 25, 2026 · 40 snapshots on file
Added new model nvidia/Muse-Spark-Open-816B-A42B-DFlash2 with other license type, available until 2026-10-16.
- Model: nvidia/Muse-Spark-Open-816B-A42B-DFlash2
- License: other
- Availability date: 2026-10-16
Full history & current snapshot →View raw diff +1 −0
+ nvidia/Muse-Spark-Open-816B-A42B-DFlash2 license:other 2026-10-16 -
Tracking since Jun 25, 2026 · 66 snapshots on file
Fast mode rate limit availability for Claude Opus 5.5 narrowed from every usage tier to Start, Build, and Scale tiers only; Evaluation tier users can access it after tier upgrade.
Full history & current snapshot →View raw diff +1 −1
- When using fast mode with speed: "fast" on Claude Opus 5.5, Claude Opus 5, or Opus 4.8, dedicated rate limits apply that are separate from standard Opus rate limits. Claude Opus 5.5 has a default fast mode rate limit in every usage tier, so you don't need to request access to use fast mode with it. Fast mode for Claude Opus 5 and Claude Opus 4.8 is in research preview. To see your fast mode rate limits, go to the Rate limits page in the Claude Console. When fast mode rate limits are exceeded, the API returns a 429 error with a retry-after header. Fast mode is not available on Claude Opus 4.7 (requests return an error) or Claude Opus 4.6. See Fast mode. + When using fast mode with speed: "fast" on Claude Opus 5.5, Claude Opus 5, or Claude Opus 4.8, dedicated rate limits apply that are separate from standard Opus rate limits. Claude Opus 5.5 has a default fast mode rate limit in the Start, Build, and Scale tiers, so you don't need to request access to use fast mode with it. Organizations in the Evaluation tier can use fast mode with Claude Opus 5.5 once they move to the Start, Build, or Scale tier. Fast mode for Claude Opus 5 and Claude Opus 4.8 is in research preview. To see your fast mode rate limits, go to the Rate limits page in the Claude Console. When fast mode rate limits are exceeded, the API returns a 429 error with a retry-after header.
-
Tracking since Jun 25, 2026 · 8 snapshots on file
Custom seat limits removed from Enterprise plan
Removed 'Custom seat limits' feature from Enterprise planFull history & current snapshot →View raw diff +0 −1
- Custom seat limits -
Tracking since Jun 25, 2026 · 11 snapshots on file
Two new Granite Switch 4.1 model variants added (30b and 8b) with Apache 2.0 license and October 2026 dates.
- granite-switch-4.1-30b
- granite-switch-4.1-8b
- license:apache-2.0
Full history & current snapshot →View raw diff +2 −0
+ ibm-granite/granite-switch-4.1-30b license:apache-2.0 2026-10-14 + ibm-granite/granite-switch-4.1-8b license:apache-2.0 2026-10-14
-
NVIDIA Nemotron models changed
Tracking since Jun 25, 2026 · 40 snapshots on file
Two NVIDIA model entries removed from the list.
- nvidia/Kimi-K2.6-DFlash removed with license:other and date 2026-06-23
- nvidia/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-BF16 removed with license:other and date 2026-06-24
Full history & current snapshot →View raw diff +0 −2
- nvidia/Kimi-K2.6-DFlash license:other 2026-06-23 - nvidia/NVIDIA-Nemotron-Labs-3-Puzzle-75B-A9B-BF16 license:other 2026-06-24
-
Tracking since Jun 25, 2026 · 20 snapshots on file
Removed microsoft/Dayhoff-3b-GR-HM-31000 model with MIT license from available models.
- microsoft/Dayhoff-3b-GR-HM-31000 model removed
- MIT license
- 2026-01-24
Full history & current snapshot →View raw diff +0 −1
- microsoft/Dayhoff-3b-GR-HM-31000 license:mit 2026-01-24 -
Tracking since Jun 25, 2026 · 9 snapshots on file
Gemini API rate limits now emphasize spend-based controls ($10–$200 per 10 min) across tiers; model-specific RPM/TPM/RPD available only in AI Studio dashboard.
Full history & current snapshot →View raw diff +11 −5
- usage tier. These limits are evaluated on a rolling 10-minute - window. Whether these limits apply to your account depends on your billing - history and account standing. - Once your project meets the specified criteria, it will be - effect within 10 minutes. Navigate to the Projects page in AI Studio to check your tiers. + usage tier. These limits are + evaluated on a rolling 10-minute window. Whether these limits apply to your + account depends on your billing history and account standing. + | Gemini 4 Argon | 5,000,000 | + | Gemini 4 Argon | 500,000,000 | + | Gemini 4 Argon | 1,000,000,000 | + Once your project meets the + specified criteria, it will be + effect within 10 minutes. Navigate to the + Projects page in AI Studio to check your + tiers.
-
Tracking since Jun 25, 2026 · 66 snapshots on file
Claude Opus 5.5 fast mode now has default rate limits in every usage tier (no access request needed); fast mode status for other models clarified as research preview.
Full history & current snapshot →View raw diff +1 −1
- When using fast mode (research preview) with speed: "fast" on Claude Opus 5.5, Claude Opus 5, or Opus 4.8, dedicated rate limits apply that are separate from standard Opus rate limits. When fast mode rate limits are exceeded, the API returns a 429 error with a retry-after header. Fast mode is not available on Claude Opus 4.7 (requests return an error) or Claude Opus 4.6 (requests to claude-opus-4-6 with speed: "fast" run at standard speed). See Fast mode. + When using fast mode with speed: "fast" on Claude Opus 5.5, Claude Opus 5, or Opus 4.8, dedicated rate limits apply that are separate from standard Opus rate limits. Claude Opus 5.5 has a default fast mode rate limit in every usage tier, so you don't need to request access to use fast mode with it. Fast mode for Claude Opus 5 and Claude Opus 4.8 is in research preview. To see your fast mode rate limits, go to the Rate limits page in the Claude Console. When fast mode rate limits are exceeded, the API returns a 429 error with a retry-after header. Fast mode is not available on Claude Opus 4.7 (requests return an error) or Claude Opus 4.6. See Fast mode.
-
Tracking since Jun 25, 2026 · 313 snapshots on file
Rosalind Academic Plan added for life sciences research; no changes to Business/Enterprise pricing.
Rosalind Academic Plan introduced for life sciences researchFull history & current snapshot →View raw diff +2 −2
- Not sure whether Business or Enterprise fits? - Ask about seats, security, deployment, or purchasing. + For life sciences research, explore the Rosalind Academic Plan. + Looking for personal plans?
-
Upcoming deprecation of selected GitHub Copilot models in mid-November
We will deprecate the following models across all GitHub Copilot experiences (including Copilot Chat, inline edits, ask and agent modes, and code completions) on November 16th, 2026: Model Deprecation date… The post Upcoming deprecation of selected GitHub Copilot models in mid-November appeared first on The GitHub Blog.
Full history & current snapshot → -
NVIDIA Nemotron models changed
Tracking since Jun 25, 2026 · 40 snapshots on file
NVIDIA Nemotron model catalog updated with four new specialized model variants including Bolt-74B and Super-VL-120B series replacing MiniMax and Privasis models.
- Added NVIDIA-Nemotron-3.5-Bolt-74B-A9B with three format variants (24GB, LiLiCorr, BF16)
- Added NVIDIA-Nemotron-3.5-Super-VL-120B-A12B with two variants (BF16, Base-BF16)
- Removed MiniMax-M2.7-DFlash and both Privasis-Cleaner models
Full history & current snapshot →View raw diff +5 −3
- nvidia/MiniMax-M2.7-DFlash license:other 2026-06-16 - nvidia/Privasis-Cleaner-0.6B license:other 2026-06-08 - nvidia/Privasis-Cleaner-4B license:other 2026-06-08 + nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-24GB license:other 2026-10-07 + nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-24GB-LiLiCorr license:other 2026-10-12 + nvidia/NVIDIA-Nemotron-3.5-Bolt-74B-A9B-BF16 license:other 2026-10-09 + nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-BF16 license:other 2026-10-12 + nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-Base-BF16 license:other 2026-10-12
-
Mistral models & deprecations changed
Tracking since Jun 25, 2026 · 18 snapshots on file
Introduced Timestral 1.0 time series forecasting models in three sizes: 123M, 28M, and 5M parameters.
- Timestral 1.0 123M - time series foundation model for zero-shot forecasting with covariate support
- Timestral 1.0 28M - time series foundation model for zero-shot forecasting with covariate support
- Timestral 1.0 5M - time series foundation model for zero-shot forecasting with covariate support
Full history & current snapshot →View raw diff +11 −0
+ Time series models + Copy section linkTime series models + Models for zero-shot time series forecasting with covariate support and probabilistic outputs. + Timestral 1.0 + A time series foundation model for zero-shot forecasting with covariate support and probabilistic outputs. + Timestral 1.0 123M + A time series foundation model for zero-shot forecasting with covariate support and probabilistic outputs. + Timestral 1.0 28M + A time series foundation model for zero-shot forecasting with covariate support and probabilistic outputs. + Timestral 1.0 5M + A time series foundation model for zero-shot forecasting with covariate support and probabilistic outputs.
-
Email notifications for repository push events have been retired
As of October 15, 2026, email notifications for repository push events have been retired and are no longer sent. We shut down this feature to help prevent spam and abuse.… The post Email notifications for repository push events have been retired appeared first on The GitHub Blog.
Full history & current snapshot → -
NVIDIA Nemotron models changed
Tracking since Jun 25, 2026 · 40 snapshots on file
Four new models added to NVIDIA Nemotron: MiniMax-M2.7-DFlash, PixelUMM, Privasis-Cleaner-0.6B, and Privasis-Cleaner-4B with other license type.
- nvidia/MiniMax-M2.7-DFlash with 2026-06-16 availability
- nvidia/PixelUMM with 2026-10-01 availability
- nvidia/Privasis-Cleaner-0.6B and nvidia/Privasis-Cleaner-4B both with 2026-06-08 availability
Full history & current snapshot →View raw diff +4 −0
+ nvidia/MiniMax-M2.7-DFlash license:other 2026-06-16 + nvidia/PixelUMM license:other 2026-10-01 + nvidia/Privasis-Cleaner-0.6B license:other 2026-06-08 + nvidia/Privasis-Cleaner-4B license:other 2026-06-08
-
Mistral AI reported benchmarks updated
Tracking since Jun 25, 2026 · 17 snapshots on file
Mistral Large 3: 1 benchmark claims (via web search)
Full history & current snapshot → -
Tracking since Jun 25, 2026 · 67 snapshots on file
GitHub API: 2 additive changes
Full history & current snapshot →View raw diff +2 −0
+ New optional parameter: include_metadata (query) + New optional parameter: include_metadata (query)
-
Tracking since Jun 25, 2026 · 25 snapshots on file
Audio translations endpoint path changed from /v1/audio/translations to v1/realtime/translations effective Feb 26, 2027.
- /v1/audio/translations renamed to v1/realtime/translations
- Change effective Feb 26, 2027
Full history & current snapshot →View raw diff +1 −0
+ | Feb 26, 2027 | /v1/audio/translations | v1/realtime/translations | -
Tracking since Jun 25, 2026 · 20 snapshots on file
Phi model changed from Dayhoff-3b-GR-HM-21000 to UserRepGen-4B with license expiration updated from 2026-01-24 to 2026-10-14.
- Model name changed: Dayhoff-3b-GR-HM-21000 to UserRepGen-4B
- License expiration date changed from 2026-01-24 to 2026-10-14
- License type remains MIT
Full history & current snapshot →View raw diff +1 −1
- microsoft/Dayhoff-3b-GR-HM-21000 license:mit 2026-01-24 + microsoft/UserRepGen-4B license:mit 2026-10-14
-
Tracking since Jun 25, 2026 · 66 snapshots on file
Documentation updated with clearer navigation paths to Billing page and Usage dashboard; no changes to rate limit values themselves.
Full history & current snapshot →View raw diff +6 −4
- Limits are set at the organization level. You can see your organization's tier and current limits on the Rate limits page in the Claude Console. - Each of the Start, Build, and Scale tiers carries a monthly spend cap, which is the maximum your organization can spend on the API each calendar month. You can view your organization's monthly spend cap and set your own limit on the Billing page. - To make the most of your rate limits, cache repeated content such as system instructions and prompts, large context documents, tool definitions, and conversation history; see prompt caching for guidance. With effective caching, you can substantially increase your actual throughput without raising your rate limits. Monitor your cache hit rate on the Usage page to tune your caching strategy. - To request higher rate limits or a higher monthly spend cap, use Request tier increase on the Rate limits page. Anthropic support can also raise limits; for urgent needs, contact Anthropic support. + Limits are set at the organization level. You can see your organization's tier and current limits on the Rate limits page (platform.claude.com/usage/limits) in the Claude Console. + Each of the Start, Build, and Scale tiers carries a monthly spend cap, which is the maximum your organization can spend on the API each calendar month. You can view your organization's monthly spend cap and set your own limit on the Billing page in the Claude Console. + ### Navigate to the Billing page + In the Claude Console, go to the Billing page (platform.claude.com/settings/billing). + To make the most of your rate limits, cache repeated content such as system instructions and prompts, large context documents, tool definitions, and conversation history; see prompt caching for guidance. With effective caching, you can substantially increase your actual throughput without raising your rate limits. Monitor your cache hit rate on the Usage page in the Claude Console to tune your caching strategy. + To request higher rate limits or a higher monthly spend cap, use Request tier increase on the Rate limits page in the Claude Console. Anthropic support can also raise limits; for urgent needs, contact Anthropic support.
-
Google Gemini API changelog changed
Tracking since Jun 25, 2026 · 29 snapshots on file
Documentation links updated from generic 'Latest model' references to specific Gemini 3.8 Flash and Gemini 3.6 Flash model guides.
- Links changed from 'Latest model guide' to 'Gemini 3.8 Flash guide' and 'Gemini 3.6 Flash guide'
- URL destinations changed from /latest-model endpoints to /whats-new-gemini-3.8 and /whats-new-gemini-3.6
Full history & current snapshot →View raw diff +4 −3
- the Latest model guide. - To learn more, see the Latest Gemini model - Latest Gemini Model + the Gemini 3.8 Flash guide. + To learn more, see the + Gemini 3.6 Flash guide. + Gemini 3.6 Flash guide
-
Google Gemini API pricing changed
Tracking since Jun 25, 2026 · 30 snapshots on file
Gemini 1.5 Flash context caching prices cut in half
Gemini 1.5 Flash · context_caching_price$0.20, prompts ≤200k; $0.40, prompts >200k→$0.10, prompts ≤200k; $0.20, prompts >200kFull history & current snapshot →View raw diff +4 −0
+ $0.015 + $9.00, prompts > 200k tokens + $0.10, prompts <= 200k tokens + $0.20, prompts > 200k tokens
-
Tracking since Jun 25, 2026 · 16 snapshots on file
Enterprise plan adds 20 seat minimum requirement
Enterprise · rate_limitUS$20/seat/month, billed annually→US$20/seat/month, billed annually, 20 seat minimumFull history & current snapshot →View raw diff +1 −1
- US$20/seat/month, billed annually. Usage cost scales with model and task. + US$20/seat/month, billed annually, 20 seat minimum. Usage cost scales with model and task.
-
Tracking since Jun 25, 2026 · 7 snapshots on file
Removed MolmoWeb-Pretrained-8B model with Apache 2.0 license from available models.
- allenai/MolmoWeb-Pretrained-8B
- license:apache-2.0
- 2026-04-09
Full history & current snapshot →View raw diff +0 −1
- allenai/MolmoWeb-Pretrained-8B license:apache-2.0 2026-04-09 -
xAI (Grok) rate limits changed
Tracking since Jun 25, 2026 · 18 snapshots on file
Voice model renamed from grok-voice-think-fast-2.0 to grok-voice-3.0 with same RPS/concurrent session limits across all tiers.
Full history & current snapshot →View raw diff +1 −1
- | grok-voice-think-fast-2.0 | T 0 T 1 T 2 T 3 T 4 | ————— | 10 20 50 100 200 | + | grok-voice-3.0 | T 0 T 1 T 2 T 3 T 4 | ————— | 10 20 50 100 200 |
-
qwen3.6-35b-a3b: Terminal-Bench accuracy (0–1) ↑ 0.23 → 0.246
Tracking since Jun 25, 2026 · 4 snapshots on file
Same model id, score moved on Terminal-Bench (Epoch) — a silent re-evaluation or model swap.
Full history & current snapshot → -
Terminal-Bench (Epoch) scores changed
Tracking since Jun 25, 2026 · 4 snapshots on file
Terminal-Bench accuracy (0–1) updated.
Full history & current snapshot →View raw diff +1 −1
- qwen3.6-35b-a3b 0.23 + qwen3.6-35b-a3b 0.246
-
Humanity's Last Exam (Epoch) scores changed
Tracking since Jun 25, 2026 · 5 snapshots on file
HLE accuracy (0–1) updated.
Full history & current snapshot →View raw diff +2 −0
+ claude-haiku-5-5_unknown 0.301 + claude-sonnet-5-5_xhigh 0.461
-
Tracking since Jun 25, 2026 · 30 snapshots on file
Arena Elo (text, overall) updated.
Full history & current snapshot →View raw diff +29 −29
- chatgpt-4o-latest-20250326 1429 - claude-opus-4-6-high 1503 - claude-opus-5-high 1503 - claude-opus-5.5-high 1515 - claude-sonnet-4-5-20250929 1439 - deepseek-r1-0528 1428 - deepseek-v4-pro-high-20260813 1446 - deepseek-v4.1-flash-max 1462 - ernie-5.0-preview-1203 1443 - gemini-3.1-pro-preview 1481 - gemini-3.5-flash-medium 1477 - gemini-3.7-flash-high 1486 - gemini-4-argon-high 1534 - glm-5.1 1461 - glm-5v-turbo 1437 - gpt-5.2-chat-latest-20260210 1439 - gpt-5.5-high 1472 - gpt-5.6-luna-xhigh 1430 - gpt-6-astra-max 1440 - hy3 1439 - mimo-v2.6-pro 1492 - mistral-large-4 1427 - muse-spark-1.1 1479 - muse-spark-1.2 (xHigh) 1482 - nvidia-nemotron-3-ultra-550b-a55b-nvfp4 1445 - qwen3.6-plus 1437 - qwen3.7-max-preview 1476 - qwen3.8-max 1483 - Step 5 Preview 1442 + chatgpt-4o-latest-20250326 1428 + claude-opus-4-6-high 1502 + claude-opus-5-high 1502 + claude-opus-5.5-high 1516 + claude-sonnet-4-5-20250929 1438 + deepseek-v4-pro-high-20260813 1445 + deepseek-v4.1-flash-max 1463 + ernie-5.0-preview-1203 1442 + gemini-3.1-pro-preview 1480 + gemini-3.5-flash-medium 1478 + gemini-3.7-flash-high 1487 + gemini-4-argon-high 1533 + glm-5.1 1463 + glm-5v-turbo 1438 + gpt-5.2-chat-latest-20260210 1438 + gpt-5.5-high 1471 + gpt-5.6-luna-xhigh 1431 + gpt-6-astra-max 1442 + hy3 1437 + mimo-v2.6-pro 1490 + mistral-large-4 1430 + muse-spark-1.1 1478 + muse-spark-1.2 (xHigh) 1485 + muse-wonder-max 1504 + nvidia-nemotron-3-ultra-550b-a55b-nvfp4 1444 + qwen3.6-plus 1436 + qwen3.7-max-preview 1475 + qwen3.8-max 1482 + Step 5 Preview 1440
-
Tracking since Jun 25, 2026 · 66 snapshots on file
Documentation clarified that not every 429 error includes retry-after header (spend-cap 429 never does), and emphasized per-model separate rate limit enforcement with shared queue visibility.
Full history & current snapshot →View raw diff +3 −3
- The rate limits for the Messages API are measured in requests per minute (RPM), input tokens per minute (ITPM), and output tokens per minute (OTPM) for each model class. If you exceed any of the rate limits you will get a 429 error describing which rate limit was exceeded, along with a retry-after header indicating how long to wait. - Rate limits are applied separately for each model; therefore you can use different models up to their respective limits simultaneously. You can check your current rate limits and behavior on the Rate limits page in the Claude Console, or read the configured limits programmatically with the Rate Limits API. - | retry-after | The number of seconds to wait until you can retry the request. Earlier retries will fail. Not sent with the spend-cap 429 (see Reaching your spend cap ). | + The rate limits for the Messages API are measured in requests per minute (RPM), input tokens per minute (ITPM), and output tokens per minute (OTPM) for each model class. If you exceed any of the rate limits you will get a 429 error describing which rate limit was exceeded. When the response has a retry-after header, wait that many seconds before retrying. + Each row in the following tables has its own RPM, ITPM, and OTPM limits. Some rows cover several models that share one limit; see the notes under the tables. You can check your current rate limits and behavior on the Rate limits page in the Claude Console, or read the configured limits programmatically with the Rate Limits API. + | retry-after | The number of seconds to wait until you can retry the request. Earlier retries will fail. Not every 429 carries it; the spend-cap 429 never does (see Reaching your spend cap ). |
-
GPQA Diamond (Epoch) scores changed
Tracking since Jun 25, 2026 · 35 snapshots on file
GPQA Diamond accuracy (0–1) updated.
Full history & current snapshot →View raw diff +2 −0
+ gemini-4-argon_high 0.964 + mistral-large-4-0_high 0.881
-
Tracking since Jun 25, 2026 · 27 snapshots on file
Origin App Marketplace launched with over 20 integrated apps from partners like Cloudflare, Linear, Sentry, and Socket Security, now open for public submissions.
- Over 20 apps integrated for deployments, CI, code review, security, monitoring, and project management
- Partners include Cloudflare, Linear, Sentry, Socket Security, Netlify, Render, Vercel, Buildkite, CircleCI, and others
- Marketplace now open to public submissions for any team building developer tools
Full history & current snapshot →View raw diff +42 −24
- Aug 27, 2026 · Changelog - Start from scratch, without a repo - Cloud Agents no longer require a connected GitHub or other third-party SCM provider to get started. Prompt from the get-go, then save your work to a Cursor Origin repo. - Select Start from scratch in the repo picker, then prompt the agent. In the background, Cursor creates an Origin repo for you. - #Turn it into a real repo, whenever you want - Once the agent's build is to your liking, click the Create repo button to save your work in an Origin repo. Choose a custom name or pick one of the suggested ones, then set the visibility to private or internal. You'll get a fully scaffolded Origin repo, ready to share or keep building on. Navigate to the Codebase tab to find your repo and access your project. - #A live preview, right in the browser - Cursor now port-forwards your cloud agent's live environment straight to your browser, so you can preview it and use tools like design mode. - #Publish your work - Connect a Vercel account and hit publish to get a live URL for what you built. - A Vercel account is required to use the publish feature. - Get started today. - Aug 27, 2026 · Changelog - Start from scratch, without a repo - Cloud Agents no longer require a connected GitHub or other third-party SCM provider to get started. Prompt from the get-go, then save your work to a Cursor Origin repo. - Select Start from scratch in the repo picker, then prompt the agent. In the background, Cursor creates an Origin repo for you. - #Turn it into a real repo, whenever you want - Once the agent's build is to your liking, click the Create repo button to save your work in an Origin repo. Choose a custom name or pick one of the suggested ones, then set the visibility to private or internal. You'll get a fully scaffolded Origin repo, ready to share or keep building on. Navigate to the Codebase tab to find your repo and access your project. - #A live preview, right in the browser - Cursor now port-forwards your cloud agent's live environment straight to your browser, so you can preview it and use tools like design mode. - #Publish your work - Connect a Vercel account and hit publish to get a live URL for what you built. - A Vercel account is required to use the publish feature. - Get started today. + Oct 13, 2026 · Changelog + Origin App Marketplace + Origin now integrates with over 20 apps for deployments, CI, code review, security, monitoring, and project management, from partners including Cloudflare, Linear, Sentry, and Socket Security. + Starting today, the marketplace is also open to public submissions, so any team building developer tools can publish an app for Origin. + #Apps for your stack + The marketplace now includes apps for: + Deployments: Cloudflare, Netlify, Render, Spacelift, and Vercel + CI: Buildkite, CircleCI, CloudBees, Depot, EngFlow, Namespace, and RWX + Code review: CodeRabbit, Greptile, and Tenki + Security: Artemis Security, Semgrep, Socket Security, and Upwind + Monitoring: Sentry + Project management: Linear + #Publish to Marketplace + To submit an app: + Create an app in your namespace. It stays private to your namespace while you build. + Build against the Origin API. Apps authenticate with an app JWT and installation access tokens, receive webhook events, and write back check runs, pull request comments, and reviews. + Make the app public, then submit it from the app's page with a slug, suggested tags, and a demo video. The video should show install, uninstall, and your core product flows. + We review the submission and list the app once it's approved. + Try the apps in the Origin App Marketplace today. + Want to build an Origin app? Start with the publishing guide and the Origin API reference. + The Origin Apps plugin can also answer questions from the live Origin API docs and plan a port to Origin for an app you already run. + Oct 13, 2026 · Changelog + Origin App Marketplace + Origin now integrates with over 20 apps for deployments, CI, code review, security, monitoring, and project management, from partners including Cloudflare, Linear, Sentry, and Socket Security. + Starting today, the marketplace is also open to public submissions, so any team building developer tools can publish an app for Origin. + #Apps for your stack + The marketplace now includes apps for: + Deployments: Cloudflare, Netlify, Render, Spacelift, and Vercel + CI: Buildkite, CircleCI, CloudBees, Depot, EngFlow, Namespace, and RWX + Code review: CodeRabbit, Greptile, and Tenki + Security: Artemis Security, Semgrep, Socket Security, and Upwind + Monitoring: Sentry + Project management: Linear + #Publish to Marketplace + To submit an app: + Create an app in your namespace. It stays private to your namespace while you build.
-
Google Gemini API changelog changed
Tracking since Jun 25, 2026 · 29 snapshots on file
Gemini 4 Argon (gemini-4-argon) released as generally available on October 13, 2026, alongside updated documentation links for deprecated Flash models.
- Gemini 4 Argon (gemini-4-argon) released for complex coding, agentic, math, and science tasks
- gemini-3.7-flash deprecated and routed to gemini-3.8-flash
- gemini-3.5-flash deprecated and routed to gemini-3.6-flash
Full history & current snapshot →View raw diff +9 −2
- Gemini 3.7 Flash deprecation: gemini-3.7-flash is deprecated and has been replaced by gemini-3.8-flash. It is no longer available to select in Google AI Studio, and API requests to gemini-3.7-flash are automatically routed to gemini-3.8-flash. Update your model string to gemini-3.8-flash and review the Gemini 3.8 Flash guide. - Gemini 3.5 Flash deprecation: gemini-3.5-flash is deprecated and has been replaced by gemini-3.6-flash. It is no longer available to select in Google AI Studio, and API requests to gemini-3.5-flash are automatically routed to gemini-3.6-flash. Update your model string to gemini-3.6-flash (or gemini-3.8-flash ) and review the Gemini 3.6 Flash guide. + October 13, 2026 + Gemini 4 Argon (gemini-4-argon) generally available (GA): Released + Gemini 4 Argon (gemini-4-argon), + our most capable model built for complex coding, agentic, math, and science + tasks. + For API changes, pricing, and migration guidance, see + What's new in Gemini 4 Argon. + Gemini 3.7 Flash deprecation: gemini-3.7-flash is deprecated and has been replaced by gemini-3.8-flash. It is no longer available to select in Google AI Studio, and API requests to gemini-3.7-flash are automatically routed to gemini-3.8-flash. Update your model string to gemini-3.8-flash and review the Gemini 3.8 Flash guide. + Gemini 3.5 Flash deprecation: gemini-3.5-flash is deprecated and has been replaced by gemini-3.6-flash. It is no longer available to select in Google AI Studio, and API requests to gemini-3.5-flash are automatically routed to gemini-3.6-flash. Update your model string to gemini-3.6-flash (or gemini-3.8-flash ) and review the Gemini 3.6 Flash guide.
-
Tracking since Jun 25, 2026 · 25 snapshots on file
Multiple GPT-4 series models including GPT-4o and GPT-4.1 variants will be removed from the API on January 14, 2027; developers can monitor shutdown dates via the List models endpoint's shutdown_date field.
- GPT-4o (gpt-4o-2024-08-06, gpt-4o-2024-11-20) removal scheduled for January 14, 2027
- GPT-4o mini (gpt-4o-mini-2024-07-18) removal scheduled for January 14, 2027
- GPT-4.1 and GPT-4.1 mini models removal scheduled for January 14, 2027
Full history & current snapshot →View raw diff +8 −0
+ To monitor model shutdowns programmatically, call the List models endpoint (GET /v1/models) and check each model’s optional shutdown_date field. When present, this field contains the shutdown date as a YYYY-MM-DD string, or null if no shutdown date has been announced; the field is also available from the Retrieve a model endpoint. + 2026-10-14: GPT-4 series model deprecations + On October 14, 2026, we notified developers using certain GPT-4 series models of their deprecation and removal from the API on January 14, 2027. + | January 14, 2027 | GPT-4o (gpt-4o-2024-08-06, gpt-4o-2024-11-20) | gpt-6-luna | + | January 14, 2027 | GPT-4o mini (gpt-4o-mini-2024-07-18) | gpt-6-luna | + | January 14, 2027 | GPT-4.1 (gpt-4.1-2025-04-14) | gpt-6-luna | + | January 14, 2027 | GPT-4.1 mini (gpt-4.1-mini-2025-04-14) | gpt-6-luna | + Fine-tuned versions of these models follow the same shutdown schedule as their underlying base models.
-
NVIDIA Nemotron models changed
Tracking since Jun 25, 2026 · 40 snapshots on file
Removed nvidia/MiniMax-M2.7-DFlash model with other license type and June 16, 2026 date.
- nvidia/MiniMax-M2.7-DFlash
- license:other
- 2026-06-16
Full history & current snapshot →View raw diff +0 −1
- nvidia/MiniMax-M2.7-DFlash license:other 2026-06-16 -
Google Gemini API pricing changed
Tracking since Jun 25, 2026 · 30 snapshots on file
Gemini 4 models added with tiered pricing; all models scheduled for 2x price increase on December 1, 2026
Gemini 4 Argon introduced with prices $2-4/1M input, $10-20/1M output (scheduled increase Dec 1, 2026)Gemini 4 Mega prices: $1-2/1M input, $5-10/1M output (scheduled increase Dec 1, 2026)Gemini 4 Flash prices: $1-2/1M input, $5-10/1M output (scheduled increase Dec 1, 2026)Gemini 4 Flash Extended prices: $3.60-7.20/1M input, $18-36/1M output (scheduled increase Dec 1, 2026)Full history & current snapshot →View raw diff +26 −0
+ Our flagship reasoning model, built for complex software engineering, deep mathematical and scientific problem solving, autonomous multi-step agents, and long-decode generation up to 1M tokens. + Note: Gemini 4 Argon is not available on the Free tier or on the Google AI Pro plan. + $2.00 through November 30, 2026. + $4.00 starting December 1, 2026. + $10.00 through November 30, 2026. + $20.00 starting December 1, 2026. + $0.10 through November 30, 2026. + $0.20 starting December 1, 2026. + $1.00 through November 30, 2026. + $2.00 starting December 1, 2026. + $5.00 through November 30, 2026. + $10.00 starting December 1, 2026. + $0.05 through November 30, 2026. + $0.10 starting December 1, 2026. + $1.00 through November 30, 2026. + $2.00 starting December 1, 2026. + $5.00 through November 30, 2026. + $10.00 starting December 1, 2026. + $0.05 through November 30, 2026. + $0.10 starting December 1, 2026. + $3.60 through November 30, 2026. + $7.20 starting December 1, 2026. + $18.00 through November 30, 2026. + $36.00 starting December 1, 2026. + $0.18 through November 30, 2026. + $0.36 starting December 1, 2026.
-
Tracking since Jun 25, 2026 · 25 snapshots on file
DigitalOcean API: 54 additive changes
Full history & current snapshot →View raw diff +54 −0
+ New endpoint: GET /v2/agents/configs + New endpoint: POST /v2/agents/configs + New endpoint: GET /v2/agents/configs/{config_id} + New endpoint: DELETE /v2/agents/configs/{config_id} + New endpoint: POST /v2/agents/configs/{config_id}/clone + New endpoint: GET /v2/agents/configs/{config_id}/sessions + New endpoint: GET /v2/agents/sessions + New endpoint: POST /v2/agents/sessions + New endpoint: POST /v2/agents/sessions/policy/validate + New endpoint: GET /v2/agents/sessions/sandbox/sizes + New endpoint: GET /v2/agents/sessions/{session_id} + New endpoint: DELETE /v2/agents/sessions/{session_id} + New endpoint: PATCH /v2/agents/sessions/{session_id} + New endpoint: GET /v2/agents/sessions/{session_id}/checkpoints + New endpoint: POST /v2/agents/sessions/{session_id}/checkpoints + New endpoint: GET /v2/agents/sessions/{session_id}/checkpoints/{checkpoint_id} + New endpoint: DELETE /v2/agents/sessions/{session_id}/checkpoints/{checkpoint_id} + New endpoint: POST /v2/agents/sessions/{session_id}/checkpoints/{checkpoint_id}/rollback + New endpoint: POST /v2/agents/sessions/{session_id}/fork + New endpoint: POST /v2/agents/sessions/{session_id}/hitl/{request_id} + New endpoint: POST /v2/agents/sessions/{session_id}/input + New endpoint: POST /v2/agents/sessions/{session_id}/pause + New endpoint: POST /v2/agents/sessions/{session_id}/resume + New endpoint: POST /v2/agents/sessions/{session_id}/sandbox/exec + New endpoint: GET /v2/agents/sessions/{session_id}/workspace/download + New endpoint: POST /v2/agents/sessions/{session_id}/workspace/transfers + New endpoint: GET /v2/agents/sessions/{session_id}/workspace/transfers/{transfer_id} + New endpoint: POST /v2/agents/sessions/{session_id}/workspace/transfers/{transfer_id}/cancel + New endpoint: POST /v2/agents/sessions/{session_id}/workspace/transfers/{transfer_id}/commit + New endpoint: POST /v2/agents/sessions/{session_id}/workspace/transfers/{transfer_id}/part-upload-urls + New endpoint: POST /v2/agents/sessions/{session_id}/workspace/upload + New endpoint: GET /v2/agents/sessions/{session_id}/events + New endpoint: GET /v2/agents/sessions/{session_id}/port-forward/{port} + New endpoint: GET /v2/agents/sessions/{session_id}/ingress + New endpoint: GET /v2/agents/templates + New endpoint: POST /v2/agents/templates + New endpoint: GET /v2/agents/templates/{template_id} + New endpoint: PUT /v2/agents/templates/{template_id} + New endpoint: DELETE /v2/agents/templates/{template_id} + New endpoint: GET /v2/agents/templates/{template_id}/builds + New endpoint: GET /v2/agents/templates/{template_id}/builds/{build_id} + New endpoint: GET /v2/agents/triggers + New endpoint: POST /v2/agents/triggers + New endpoint: GET /v2/agents/triggers/by-session/{session_id} + New endpoint: GET /v2/agents/triggers/reusable-sessions + New endpoint: GET /v2/agents/triggers/{trigger_id} + New endpoint: DELETE /v2/agents/triggers/{trigger_id} + New endpoint: PATCH /v2/agents/triggers/{trigger_id} + New endpoint: GET /v2/agents/triggers/{trigger_id}/executions + New endpoint: GET /v2/agents/triggers/{trigger_id}/executions/{execution_id} + New endpoint: POST /v2/agents/triggers/{trigger_id}/executions/{execution_id}/cancel + New endpoint: POST /v2/agents/triggers/{trigger_id}/rotate-secret + New endpoint: POST /v2/agents/triggers/{trigger_id}/webhook + New endpoint: GET /v2/agents/webhook-providers
-
Tracking since Jun 25, 2026 · 313 snapshots on file
API pricing details removed; page now focuses on ChatGPT Business/Enterprise subscriptions rather than developer API rates.
Multiple API pricing rates removed from page (GPT-4 Turbo, GPT-4, GPT-3.5, Vision, Audio, Real-time, Search, Storage pricing no longer listed)Full history & current snapshot →View raw diff +213 −43
- Business Pricing | OpenAI - Price - Input:$10.00 / 1M tokens Cached input:$1.00 / 1M tokens Output:$50.00 / 1M tokens - Price - Input:$2.00 / 1M tokens Cached input:$0.10 / 1M tokens Output:$10.00 / 1M tokens - Price - Input:$0.10 / 1M tokens Cached input:$0.01 / 1M tokens Output:$0.50 / 1M tokens - Price - Image:$8.00 / 1M tokens for inputs$2.00 / 1M tokens for cached inputs$30.00 / 1M tokens for outputs Text:$5.00 / 1M tokens for inputs$1.25 / 1M tokens for cached inputs - Price - Image:$8.00 / 1M tokens for inputs$2.00 / 1M tokens for cached inputs$30.00 / 1M tokens for outputs Text:$5.00 / 1M tokens for inputs$1.25 / 1M tokens for cached inputs - Price - $0.05 per minute / $0.00083 per second - Price - Audio:$32.00 / 1M tokens for inputs$0.40 / 1M tokens for cached inputs$64.00 / 1M tokens for outputs Text:$4.00 / 1M tokens for inputs$0.40 / 1M tokens for cached inputs$24.00 / 1M tokens for outputs Image:$5.00 / 1M tokens for inputs$0.50 / 1M tokens for cached inputs - Price - Audio:$10.00 / 1M tokens for inputs$0.30 / 1M tokens for cached inputs$20.00 / 1M tokens for outputs Text:$0.60 / 1M tokens for inputs$0.06 / 1M tokens for cached inputs$2.40 / 1M tokens for outputs Image:$0.80 / 1M tokens for inputs$0.08 / 1M tokens for cached inputs - Price - $0.017 per minute / $0.00028 per second - Price - $0.0045 per minute / $0.00008 per second - Price - $0.034 per minute / $0.00057 per second - Price - $10.00 / 1k calls Search content tokens are free. - Price - Now:1 GB for $0.03 / 64GB for $1.92 per container Starting March 31, 2026:1 GB for $0.03 / 64GB for $1.92 per 20-minute session per container - Provides lower costs for requests in exchange for slower response times and occasional resource unavailability. Ideal for non-production or lower priority tasks. - Enterprise offerings - Contact our sales team to learn more about Data residency(opens in a new window), Scale Tier andReserved Capacity designed for cutting-edge customers running larger workloads. - We recommend experimenting with all of these models in the Playground(opens in a new window) to explore which models provide the best price performance trade-off for your usage. - Do you offer an enterprise package or SLAs? - We offer different tiers of access to our enterprise customers that include SLAs, lower latency, and more. Please contact our sales team to learn more. - Yes, we treat Playground usage the same as regular API usage. You will be billed at the per-token input and output prices mentioned above. - How will I know how many tokens I’ve used each month? - A token is a mathematical representation of natural language. Log in to your account to view your usage tracking dashboard(opens in a new window). This dashboard will show you how many tokens you’ve used during the current and past billing cycles. - You can set a monthly budget in your billing settings(opens in a new window), after which we’ll stop serving your requests. There may be a delay in enforcing the limit, and you are responsible for any overage incurred. You can also configure an email notification threshold to receive an email alert once you cross that threshold each month. We recommend checking your usage tracking dashboard(opens in a new window) regularly to monitor your spend. - Is access to the API included in ChatGPT Plus, Business, Enterprise or Edu? - No, OpenAI APIs are billed separately from ChatGPT Plus, Business, Enterprise and Edu. ChatGPT subscription pricing can be found at openai.com/chatgpt/pricing/. - Images are converted into tokens and charged per token. Text models price image tokens at standard text token rates, while GPT Image and gpt-realtime uses a separate image token rate. Models like gpt-4.1-mini, gpt-4.1-nano, and o4-mini convert images into tokens differently. Learn more in our docs(opens in a new window). - ChatGPT Business(opens in a new window) - ChatGPT Enterprise(opens in a new window) - ChatGPT for Education(opens in a new window) + ChatGPT Business(opens in a new window) + ChatGPT Enterprise(opens in a new window) + ChatGPT for Education(opens in a new window) + Image 1 + Business + A secure workspace with company context and flexible seat types for any budget. + Get started(opens in a new window) + Standard seat + $20/ month + $20/month if billed annually. $25/month if billed monthly. + Premium seat + $100/ month + 5x more usage than standard, with no 5-hour limit + $100/month if billed annually. $125/month if billed monthly. + No training on your business data by default + Mix and match seat types + For teams of 2–200 employees. Unlimited subject to abuse guardrails. Learn more(opens in a new window)
-
Tracking since Jun 25, 2026 · 67 snapshots on file
GitHub API: 10 additive changes
Full history & current snapshot →View raw diff +10 −0
+ New endpoint: GET /agents/automations + New endpoint: GET /agents/automations/available-tools + New endpoint: GET /agents/automations/available-triggers + New endpoint: GET /agents/automations/{automation_id} + New endpoint: GET /agents/automations/{automation_id}/tasks + New endpoint: GET /agents/repos/{owner}/{repo}/automations + New endpoint: POST /agents/repos/{owner}/{repo}/automations + New endpoint: GET /agents/repos/{owner}/{repo}/automations/{automation_id} + New endpoint: DELETE /agents/repos/{owner}/{repo}/automations/{automation_id} + New endpoint: PATCH /agents/repos/{owner}/{repo}/automations/{automation_id}
-
Tracking since Jun 25, 2026 · 124 snapshots on file
OpenAI API: 1 additive change
Full history & current snapshot →View raw diff +1 −0
+ New endpoint: POST /organization/projects/{project_id}/api_keys/{api_key_id} -
Tracking since Jun 25, 2026 · 313 snapshots on file
OpenAI API maintains token-based pricing with no rate changes detected
No meaningful pricing change detected - page shows API token pricing model maintained with existing rate structuresFull history & current snapshot →View raw diff +43 −213
- Image 1 - Business - A secure workspace with company context and flexible seat types for any budget. - Get started(opens in a new window) - Standard seat - NZ$33/ month - $20/month if billed annually. $25/month if billed monthly. - Premium seat - NZ$165/ month - 5x more usage than standard, with no 5-hour limit - $100/month if billed annually. $125/month if billed monthly. - No training on your business data by default - Mix and match seat types - For teams of 2–200 employees. Unlimited subject to abuse guardrails. Learn more(opens in a new window) - Image 2 - Enterprise - Enterprise-grade AI, security, and support for businesses operating at scale - Contact our sales team to discuss enterprise pricing.* - Enterprise-level security and controls, including SCIM, EKM, user analytics, domain verification, and role-based access controls - Advanced data privacy with custom data retention policies, encryption at rest and in transit, and no training on your business data by default. Learn more - *Credit-based(opens in a new window) pricing and token-based(opens in a new window) pricing are available for Enterprise plans. - Looking for personal plans? - Compare features across plans - Try ChatGPT Business(opens in a new window) - Enterprise - Unlimited*Plan: Business, Feature: Everyday text chats, Unlimited* - Unlimited*Plan: Enterprise, Feature: Everyday text chats, Unlimited* - Unlimited*Plan: Business, Feature: Chat history, Unlimited* - Unlimited*Plan: Enterprise, Feature: Chat history, Unlimited* - Plan: Business, Feature: Access on web, iOS, Android, Yes - Plan: Enterprise, Feature: Access on web, iOS, Android, Yes - FlexiblePlan: Business, Feature: GPT-6.1 Sol, Flexible - FlexiblePlan: Enterprise, Feature: GPT-6.1 Sol, Flexible - FlexiblePlan: Business, Feature: GPT-6 Astra, Flexible - FlexiblePlan: Enterprise, Feature: GPT-6 Astra, Flexible - FlexiblePlan: Business, Feature: GPT-6 Sol, Flexible - FlexiblePlan: Enterprise, Feature: GPT-6 Sol, Flexible - FlexiblePlan: Business, Feature: GPT-6 Luna, Flexible - FlexiblePlan: Enterprise, Feature: GPT-6 Luna, Flexible - FlexiblePlan: Business, Feature: GPT-5.6 Sol, Flexible - FlexiblePlan: Enterprise, Feature: GPT-5.6 Sol, Flexible - GPT-5.6 Sol Pro - FlexiblePlan: Business, Feature: GPT-5.6 Sol Pro, Flexible - FlexiblePlan: Enterprise, Feature: GPT-5.6 Sol Pro, Flexible - FlexiblePlan: Business, Feature: GPT-5.6 Terra, Flexible - FlexiblePlan: Enterprise, Feature: GPT-5.6 Terra, Flexible - FlexiblePlan: Business, Feature: GPT-5.6 Luna, Flexible - FlexiblePlan: Enterprise, Feature: GPT-5.6 Luna, Flexible - FlexiblePlan: Business, Feature: GPT-5 Thinking Mini, Flexible - FlexiblePlan: Enterprise, Feature: GPT-5 Thinking Mini, Flexible - Plan: Business, Feature: Legacy models, Yes - Plan: Enterprise, Feature: Legacy models, Yes - Fast Plan: Business, Feature: Response times, Fast - Fastest Plan: Enterprise, Feature: Response times, Fastest - 54K Plan: Business, Feature: GPT Instant total context window, 54K - 128K Plan: Enterprise, Feature: GPT Instant total context window, 128K - ~40 pages Plan: Business, Feature: GPT Instant input maximum***, ~40 pages - ~250 pages Plan: Enterprise, Feature: GPT Instant input maximum***, ~250 pages - 256K Plan: Business, Feature: GPT Reasoning total context window, 256K - 256K Plan: Enterprise, Feature: GPT Reasoning total context window, 256K
-
Google Gemini API changelog changed
Tracking since Jun 25, 2026 · 29 snapshots on file
Deprecated Gemini models gemini-3.7-flash and gemini-3.5-flash are no longer available to select in Google AI Studio, with automatic API routing to their replacements.
- gemini-3.7-flash no longer available to select in Google AI Studio, auto-routed to gemini-3.8-flash
- gemini-3.5-flash no longer available to select in Google AI Studio, auto-routed to gemini-3.6-flash
Full history & current snapshot →View raw diff +2 −2
- Gemini 3.7 Flash deprecation: gemini-3.7-flash is deprecated and has been replaced by gemini-3.8-flash. All requests to gemini-3.7-flash are automatically routed to gemini-3.8-flash. Update your model string to gemini-3.8-flash and review the Gemini 3.8 Flash guide. - Gemini 3.5 Flash deprecation: gemini-3.5-flash is deprecated and has been replaced by gemini-3.6-flash. All requests to gemini-3.5-flash are automatically routed to gemini-3.6-flash. Update your model string to gemini-3.6-flash (or gemini-3.8-flash ) and review the Gemini 3.6 Flash guide. + Gemini 3.7 Flash deprecation: gemini-3.7-flash is deprecated and has been replaced by gemini-3.8-flash. It is no longer available to select in Google AI Studio, and API requests to gemini-3.7-flash are automatically routed to gemini-3.8-flash. Update your model string to gemini-3.8-flash and review the Gemini 3.8 Flash guide. + Gemini 3.5 Flash deprecation: gemini-3.5-flash is deprecated and has been replaced by gemini-3.6-flash. It is no longer available to select in Google AI Studio, and API requests to gemini-3.5-flash are automatically routed to gemini-3.6-flash. Update your model string to gemini-3.6-flash (or gemini-3.8-flash ) and review the Gemini 3.6 Flash guide.
-
Google Gemini deprecations changed
Tracking since Jun 25, 2026 · 25 snapshots on file
Updated deprecation status descriptions for gemini-3.7-flash and gemini-3.5-flash to clarify they are not available in Google AI Studio but continue routing requests automatically without shutdown dates.
- gemini-3.7-flash shutdown date changed from 'No shutdown date announced' to 'N/A'
- gemini-3.5-flash shutdown date changed from 'No shutdown date announced' to 'N/A'
- Both models now noted as 'not available to select in Google AI Studio' with clarification that endpoints are not shutting down
Full history & current snapshot →View raw diff +2 −2
- | gemini-3.7-flash | August 13, 2026 | No shutdown date announced | gemini-3.8-flash (requests to gemini-3.7-flash are automatically routed to gemini-3.8-flash) | - | gemini-3.5-flash | May 19, 2026 | No shutdown date announced | gemini-3.6-flash (requests to gemini-3.5-flash are automatically routed to gemini-3.6-flash) | + | gemini-3.7-flash | August 13, 2026 | N/A | gemini-3.8-flash (not available to select in Google AI Studio; the endpoint is not shutting down and automatically routes API requests to gemini-3.8-flash) | + | gemini-3.5-flash | May 19, 2026 | N/A | gemini-3.6-flash (not available to select in Google AI Studio; the endpoint is not shutting down and automatically routes API requests to gemini-3.6-flash) |
-
Tracking since Jun 25, 2026 · 19 snapshots on file
Cursor replaced Pro/Pro+ tiers with new Free/Plus/Premium/Super/Ultra plan structure
Pricing structure completely overhauled: removed Pro ($20), Pro+ ($60) plan tier and replaced with Free/Plus/Premium/Super/Ultra lineupFull history & current snapshot →View raw diff +76 −48
- IndividualTeams & Enterprise - ✓ No credit card required - ✓ Limited Agent requests - Pro - $20 / mo. - ✓ Extended limits on Agent - ✓ Generous limits for Grok - Get Pro - Pro+ - $60 / mo. - Everything in Pro, plus: - ✓ 3x Pro limits on Agent - ✓ Generous limits for Grok - ✓ Higher Grok Bot usage limits - Get Pro+ - $200 / mo. - Everything in Pro, plus: - ✓ 20x Pro limits on Agent - ✓ Generous limits for Grok - ✓ Highest Grok Bot usage limits - We recommend Pro+ for daily agent users, and Ultra for agent power users. The Teams plan is recommended for professionals collaborating with others, and larger organizations that need invoicing, pooled usage, or advanced security should choose Enterprise. - Every plan includes a set amount of model usage. On-demand usage allows you to continue using models after your included amount is consumed, billed in arrears. See our docs for more details. - All prices are exclusive of any applicable taxes. See our Terms of Service for more information. - You're welcome to join our forum and share your thoughts! If you prefer a private conversation, feel free to email us directly at hi@cursor.com. - IndividualTeams & Enterprise - ✓ No credit card required - ✓ Limited Agent requests - Pro - $20 / mo. - ✓ Extended limits on Agent - ✓ Generous limits for Grok - Get Pro - Pro+ - $60 / mo. - Everything in Pro, plus: - ✓ 3x Pro limits on Agent - ✓ Generous limits for Grok - ✓ Higher Grok Bot usage limits - Get Pro+ - $200 / mo. - Everything in Pro, plus: - ✓ 20x Pro limits on Agent - ✓ Generous limits for Grok - ✓ Highest Grok Bot usage limits - We recommend Pro+ for daily agent users, and Ultra for agent power users. The Teams plan is recommended for professionals collaborating with others, and larger organizations that need invoicing, pooled usage, or advanced security should choose Enterprise. - Every plan includes a set amount of model usage. On-demand usage allows you to continue using models after your included amount is consumed, billed in arrears. See our docs for more details. - All prices are exclusive of any applicable taxes. See our Terms of Service for more information. - You're welcome to join our forum and share your thoughts! If you prefer a private conversation, feel free to email us directly at hi@cursor.com. + $0/month + Limited + Limited agent requests + Limited Chat requests + $8/month + Our introductory plan: + *Only Grok models are available on this plan + $30/month + Third-party models at API list price + ✓Ad-free scrolling on most surfaces + $100/month + $200/month
-
Meta reported benchmarks updated
Tracking since Jun 25, 2026 · 17 snapshots on file
Llama 3.1 405B Instruct: 11 benchmark claims (via web search)
Full history & current snapshot → -
Tracking since Jun 25, 2026 · 313 snapshots on file
Business plan pricing structure clarified with explicit annual/monthly billing tiers
Business Standard seat · price$20/month→$20/month if billed annually. $25/month if billed monthly.Business Premium seat · price$100/month→$100/month if billed annually. $125/month if billed monthly.Full history & current snapshot →View raw diff +3 −4
- $20/ month - $100/ month - Not sure whether Business or Enterprise fits? - Ask about seats, security, deployment, or purchasing. + NZ$33/ month + NZ$165/ month + Looking for personal plans?
-
xAI reported benchmarks updated
Tracking since Jun 25, 2026 · 17 snapshots on file
Grok 4.7: 7 benchmark claims (via web search)
Full history & current snapshot → -
Tracking since Jun 25, 2026 · 66 snapshots on file
Removed Start tier specification; page emphasizes dashboard-only visibility of actual rate limits by usage tier.
Full history & current snapshot →View raw diff +0 −2
- Start tier - Start tier
-
Tracking since Jun 25, 2026 · 313 snapshots on file
ChatGPT Business plan introduced with Standard seat at $20/month and Premium seat at $100/month, replacing previous navigation links.
now$20/mo$25/mo$100/mo$125/mo2usersFull history & current snapshot →View raw diff +214 −3
- ChatGPT Business(opens in a new window) - ChatGPT Enterprise(opens in a new window) - ChatGPT for Education(opens in a new window) + Image 1 + Business + A secure workspace with company context and flexible seat types for any budget. + Get started(opens in a new window) + Standard seat + $20/ month + $20/month if billed annually. $25/month if billed monthly. + Premium seat + $100/ month + 5x more usage than standard, with no 5-hour limit + $100/month if billed annually. $125/month if billed monthly. + No training on your business data by default + Mix and match seat types + For teams of 2–200 employees. Unlimited subject to abuse guardrails. Learn more(opens in a new window) + Image 2 + Enterprise + Enterprise-grade AI, security, and support for businesses operating at scale + Contact our sales team to discuss enterprise pricing.* + Enterprise-level security and controls, including SCIM, EKM, user analytics, domain verification, and role-based access controls + Advanced data privacy with custom data retention policies, encryption at rest and in transit, and no training on your business data by default. Learn more + *Credit-based(opens in a new window) pricing and token-based(opens in a new window) pricing are available for Enterprise plans. + Not sure whether Business or Enterprise fits? + Ask about seats, security, deployment, or purchasing. + Compare features across plans + Try ChatGPT Business(opens in a new window) + Enterprise + Unlimited*Plan: Business, Feature: Everyday text chats, Unlimited* + Unlimited*Plan: Enterprise, Feature: Everyday text chats, Unlimited* + Unlimited*Plan: Business, Feature: Chat history, Unlimited* + Unlimited*Plan: Enterprise, Feature: Chat history, Unlimited* + Plan: Business, Feature: Access on web, iOS, Android, Yes + Plan: Enterprise, Feature: Access on web, iOS, Android, Yes + FlexiblePlan: Business, Feature: GPT-6.1 Sol, Flexible + FlexiblePlan: Enterprise, Feature: GPT-6.1 Sol, Flexible + FlexiblePlan: Business, Feature: GPT-6 Astra, Flexible + FlexiblePlan: Enterprise, Feature: GPT-6 Astra, Flexible + FlexiblePlan: Business, Feature: GPT-6 Sol, Flexible + FlexiblePlan: Enterprise, Feature: GPT-6 Sol, Flexible + FlexiblePlan: Business, Feature: GPT-6 Luna, Flexible + FlexiblePlan: Enterprise, Feature: GPT-6 Luna, Flexible + FlexiblePlan: Business, Feature: GPT-5.6 Sol, Flexible + FlexiblePlan: Enterprise, Feature: GPT-5.6 Sol, Flexible + GPT-5.6 Sol Pro + FlexiblePlan: Business, Feature: GPT-5.6 Sol Pro, Flexible + FlexiblePlan: Enterprise, Feature: GPT-5.6 Sol Pro, Flexible + FlexiblePlan: Business, Feature: GPT-5.6 Terra, Flexible + FlexiblePlan: Enterprise, Feature: GPT-5.6 Terra, Flexible + FlexiblePlan: Business, Feature: GPT-5.6 Luna, Flexible + FlexiblePlan: Enterprise, Feature: GPT-5.6 Luna, Flexible + FlexiblePlan: Business, Feature: GPT-5 Thinking Mini, Flexible + FlexiblePlan: Enterprise, Feature: GPT-5 Thinking Mini, Flexible + Plan: Business, Feature: Legacy models, Yes + Plan: Enterprise, Feature: Legacy models, Yes + Fast Plan: Business, Feature: Response times, Fast + Fastest Plan: Enterprise, Feature: Response times, Fastest + 54K Plan: Business, Feature: GPT Instant total context window, 54K + 128K Plan: Enterprise, Feature: GPT Instant total context window, 128K
-
Moonshot AI reported benchmarks updated
Tracking since Jun 25, 2026 · 17 snapshots on file
Kimi K3: 12 benchmark claims (via web search)
Full history & current snapshot → -
Zhipu AI (Z.ai) reported benchmarks updated
Tracking since Jun 25, 2026 · 17 snapshots on file
GLM-5.3: 10 benchmark claims (via web search)
Full history & current snapshot → -
Alibaba reported benchmarks updated
Tracking since Jun 25, 2026 · 17 snapshots on file
Qwen3.8-Max: 12 benchmark claims (via web search)
Full history & current snapshot → -
DeepSeek reported benchmarks updated
Tracking since Jun 25, 2026 · 17 snapshots on file
DeepSeek-V4-Pro: 10 benchmark claims (via web search)
Full history & current snapshot → -
Google reported benchmarks updated
Tracking since Jun 25, 2026 · 17 snapshots on file
Gemini 3.1 Pro: 11 benchmark claims (via web search)
Full history & current snapshot → -
OpenAI reported benchmarks updated
Tracking since Jun 25, 2026 · 17 snapshots on file
GPT-6 Astra: 10 benchmark claims (via web search)
Full history & current snapshot → -
Anthropic reported benchmarks updated
Tracking since Jun 25, 2026 · 17 snapshots on file
Claude Opus 5.5: 12 benchmark claims (via web search)
Full history & current snapshot → -
Tracking since Jun 25, 2026 · 25 snapshots on file
DigitalOcean API: 4 breaking changes · 4 additive
Full history & current snapshot →View raw diff +4 −4
- Endpoint removed: GET /v2/gen-ai/simulation_runs/{run_uuid}/journeys - Endpoint removed: GET /v2/gen-ai/simulation_runs/{run_uuid}/journeys/{journey_uuid} - Endpoint removed: GET /v2/gen-ai/simulation_runs/{run_uuid}/journeys/{journey_uuid}/trajectory - Endpoint removed: GET /v2/gen-ai/simulation_runs/{run_uuid}/journeys/{journey_uuid}/trajectory_url + New endpoint: GET /v2/gen-ai/simulation_runs/{run_uuid}/episodes + New endpoint: GET /v2/gen-ai/simulation_runs/{run_uuid}/episodes/{episode_uuid} + New endpoint: GET /v2/gen-ai/simulation_runs/{run_uuid}/episodes/{episode_uuid}/trajectory + New endpoint: GET /v2/gen-ai/simulation_runs/{run_uuid}/episodes/{episode_uuid}/trajectory_url
-
Tracking since Jun 25, 2026 · 313 snapshots on file
Page shifted from API token pricing to ChatGPT Business/Enterprise subscription plans
API pricing information removed from current page; content now focuses on ChatGPT Business/Enterprise subscription plans rather than API token ratesFull history & current snapshot →View raw diff +214 −43
- Business Pricing | OpenAI - Price - Input:$10.00 / 1M tokens Cached input:$1.00 / 1M tokens Output:$50.00 / 1M tokens - Price - Input:$2.00 / 1M tokens Cached input:$0.10 / 1M tokens Output:$10.00 / 1M tokens - Price - Input:$0.10 / 1M tokens Cached input:$0.01 / 1M tokens Output:$0.50 / 1M tokens - Price - Image:$8.00 / 1M tokens for inputs$2.00 / 1M tokens for cached inputs$30.00 / 1M tokens for outputs Text:$5.00 / 1M tokens for inputs$1.25 / 1M tokens for cached inputs - Price - Image:$8.00 / 1M tokens for inputs$2.00 / 1M tokens for cached inputs$30.00 / 1M tokens for outputs Text:$5.00 / 1M tokens for inputs$1.25 / 1M tokens for cached inputs - Price - $0.05 per minute / $0.00083 per second - Price - Audio:$32.00 / 1M tokens for inputs$0.40 / 1M tokens for cached inputs$64.00 / 1M tokens for outputs Text:$4.00 / 1M tokens for inputs$0.40 / 1M tokens for cached inputs$24.00 / 1M tokens for outputs Image:$5.00 / 1M tokens for inputs$0.50 / 1M tokens for cached inputs - Price - Audio:$10.00 / 1M tokens for inputs$0.30 / 1M tokens for cached inputs$20.00 / 1M tokens for outputs Text:$0.60 / 1M tokens for inputs$0.06 / 1M tokens for cached inputs$2.40 / 1M tokens for outputs Image:$0.80 / 1M tokens for inputs$0.08 / 1M tokens for cached inputs - Price - $0.017 per minute / $0.00028 per second - Price - $0.0045 per minute / $0.00008 per second - Price - $0.034 per minute / $0.00057 per second - Price - $10.00 / 1k calls Search content tokens are free. - Price - Now:1 GB for $0.03 / 64GB for $1.92 per container Starting March 31, 2026:1 GB for $0.03 / 64GB for $1.92 per 20-minute session per container - Provides lower costs for requests in exchange for slower response times and occasional resource unavailability. Ideal for non-production or lower priority tasks. - Enterprise offerings - Contact our sales team to learn more about Data residency(opens in a new window), Scale Tier andReserved Capacity designed for cutting-edge customers running larger workloads. - We recommend experimenting with all of these models in the Playground(opens in a new window) to explore which models provide the best price performance trade-off for your usage. - Do you offer an enterprise package or SLAs? - We offer different tiers of access to our enterprise customers that include SLAs, lower latency, and more. Please contact our sales team to learn more. - Yes, we treat Playground usage the same as regular API usage. You will be billed at the per-token input and output prices mentioned above. - How will I know how many tokens I’ve used each month? - A token is a mathematical representation of natural language. Log in to your account to view your usage tracking dashboard(opens in a new window). This dashboard will show you how many tokens you’ve used during the current and past billing cycles. - You can set a monthly budget in your billing settings(opens in a new window), after which we’ll stop serving your requests. There may be a delay in enforcing the limit, and you are responsible for any overage incurred. You can also configure an email notification threshold to receive an email alert once you cross that threshold each month. We recommend checking your usage tracking dashboard(opens in a new window) regularly to monitor your spend. - Is access to the API included in ChatGPT Plus, Business, Enterprise or Edu? - No, OpenAI APIs are billed separately from ChatGPT Plus, Business, Enterprise and Edu. ChatGPT subscription pricing can be found at openai.com/chatgpt/pricing/. - Images are converted into tokens and charged per token. Text models price image tokens at standard text token rates, while GPT Image and gpt-realtime uses a separate image token rate. Models like gpt-4.1-mini, gpt-4.1-nano, and o4-mini convert images into tokens differently. Learn more in our docs(opens in a new window). - ChatGPT Business(opens in a new window) - ChatGPT Enterprise(opens in a new window) - ChatGPT for Education(opens in a new window) + Image 1 + Business + A secure workspace with company context and flexible seat types for any budget. + Get started(opens in a new window) + Standard seat + $20/ month + $20/month if billed annually. $25/month if billed monthly. + Premium seat + $100/ month + 5x more usage than standard, with no 5-hour limit + $100/month if billed annually. $125/month if billed monthly. + No training on your business data by default + Mix and match seat types + For teams of 2–200 employees. Unlimited subject to abuse guardrails. Learn more(opens in a new window) + Image 2 + Enterprise + Enterprise-grade AI, security, and support for businesses operating at scale
-
Tracking since Jun 25, 2026 · 20 snapshots on file
Microsoft Phi model roster updated: Dayhoff-170M-UR90-61000 and Dayhoff-3b-GR-HM-11000 removed, replaced with AesCode-32B and AesCode-8B under apache-2.0 license.
- Dayhoff-170M-UR90-61000 and Dayhoff-3b-GR-HM-11000 removed
- AesCode-32B and AesCode-8B added
- License changed from mit to apache-2.0
Full history & current snapshot →View raw diff +2 −2
- microsoft/Dayhoff-170M-UR90-61000 license:mit 2026-01-24 - microsoft/Dayhoff-3b-GR-HM-11000 license:mit 2026-01-24 + microsoft/AesCode-32B license:apache-2.0 2026-09-29 + microsoft/AesCode-8B license:apache-2.0 2026-09-29
-
Tracking since Jun 25, 2026 · 313 snapshots on file
OpenAI API pricing page refocused on token-based usage rates; ChatGPT subscription plans moved to separate product page.
Removed ChatGPT Business/Enterprise subscription tiers from API pricing page focus (shifted to separate ChatGPT product page)Full history & current snapshot →View raw diff +43 −213
- Image 1 - Business - A secure workspace with company context and flexible seat types for any budget. - Get started(opens in a new window) - Standard seat - $20/ month - $20/month if billed annually. $25/month if billed monthly. - Premium seat - $100/ month - 5x more usage than standard, with no 5-hour limit - $100/month if billed annually. $125/month if billed monthly. - No training on your business data by default - Mix and match seat types - For teams of 2–200 employees. Unlimited subject to abuse guardrails. Learn more(opens in a new window) - Image 2 - Enterprise - Enterprise-grade AI, security, and support for businesses operating at scale - Contact our sales team to discuss enterprise pricing.* - Enterprise-level security and controls, including SCIM, EKM, user analytics, domain verification, and role-based access controls - Advanced data privacy with custom data retention policies, encryption at rest and in transit, and no training on your business data by default. Learn more - *Credit-based(opens in a new window) pricing and token-based(opens in a new window) pricing are available for Enterprise plans. - Looking for personal plans? - Compare features across plans - Try ChatGPT Business(opens in a new window) - Enterprise - Unlimited*Plan: Business, Feature: Everyday text chats, Unlimited* - Unlimited*Plan: Enterprise, Feature: Everyday text chats, Unlimited* - Unlimited*Plan: Business, Feature: Chat history, Unlimited* - Unlimited*Plan: Enterprise, Feature: Chat history, Unlimited* - Plan: Business, Feature: Access on web, iOS, Android, Yes - Plan: Enterprise, Feature: Access on web, iOS, Android, Yes - FlexiblePlan: Business, Feature: GPT-6.1 Sol, Flexible - FlexiblePlan: Enterprise, Feature: GPT-6.1 Sol, Flexible - FlexiblePlan: Business, Feature: GPT-6 Astra, Flexible - FlexiblePlan: Enterprise, Feature: GPT-6 Astra, Flexible - FlexiblePlan: Business, Feature: GPT-6 Sol, Flexible - FlexiblePlan: Enterprise, Feature: GPT-6 Sol, Flexible - FlexiblePlan: Business, Feature: GPT-6 Luna, Flexible - FlexiblePlan: Enterprise, Feature: GPT-6 Luna, Flexible - FlexiblePlan: Business, Feature: GPT-5.6 Sol, Flexible - FlexiblePlan: Enterprise, Feature: GPT-5.6 Sol, Flexible - GPT-5.6 Sol Pro - FlexiblePlan: Business, Feature: GPT-5.6 Sol Pro, Flexible - FlexiblePlan: Enterprise, Feature: GPT-5.6 Sol Pro, Flexible - FlexiblePlan: Business, Feature: GPT-5.6 Terra, Flexible - FlexiblePlan: Enterprise, Feature: GPT-5.6 Terra, Flexible - FlexiblePlan: Business, Feature: GPT-5.6 Luna, Flexible - FlexiblePlan: Enterprise, Feature: GPT-5.6 Luna, Flexible - FlexiblePlan: Business, Feature: GPT-5 Thinking Mini, Flexible - FlexiblePlan: Enterprise, Feature: GPT-5 Thinking Mini, Flexible - Plan: Business, Feature: Legacy models, Yes - Plan: Enterprise, Feature: Legacy models, Yes - Fast Plan: Business, Feature: Response times, Fast - Fastest Plan: Enterprise, Feature: Response times, Fastest - 54K Plan: Business, Feature: GPT Instant total context window, 54K - 128K Plan: Enterprise, Feature: GPT Instant total context window, 128K - ~40 pages Plan: Business, Feature: GPT Instant input maximum***, ~40 pages - ~250 pages Plan: Enterprise, Feature: GPT Instant input maximum***, ~250 pages - 256K Plan: Business, Feature: GPT Reasoning total context window, 256K - 256K Plan: Enterprise, Feature: GPT Reasoning total context window, 256K
-
Tracking since Jun 25, 2026 · 313 snapshots on file
OpenAI pricing page structure changed from API token-based to ChatGPT Business/Enterprise seat-based plans
Previous API pricing page removed; ChatGPT Business/Enterprise seat-based pricing now displayedFull history & current snapshot →View raw diff +213 −43
- Business Pricing | OpenAI - Price - Input:$10.00 / 1M tokens Cached input:$1.00 / 1M tokens Output:$50.00 / 1M tokens - Price - Input:$2.00 / 1M tokens Cached input:$0.10 / 1M tokens Output:$10.00 / 1M tokens - Price - Input:$0.10 / 1M tokens Cached input:$0.01 / 1M tokens Output:$0.50 / 1M tokens - Price - Image:$8.00 / 1M tokens for inputs$2.00 / 1M tokens for cached inputs$30.00 / 1M tokens for outputs Text:$5.00 / 1M tokens for inputs$1.25 / 1M tokens for cached inputs - Price - Image:$8.00 / 1M tokens for inputs$2.00 / 1M tokens for cached inputs$30.00 / 1M tokens for outputs Text:$5.00 / 1M tokens for inputs$1.25 / 1M tokens for cached inputs - Price - $0.05 per minute / $0.00083 per second - Price - Audio:$32.00 / 1M tokens for inputs$0.40 / 1M tokens for cached inputs$64.00 / 1M tokens for outputs Text:$4.00 / 1M tokens for inputs$0.40 / 1M tokens for cached inputs$24.00 / 1M tokens for outputs Image:$5.00 / 1M tokens for inputs$0.50 / 1M tokens for cached inputs - Price - Audio:$10.00 / 1M tokens for inputs$0.30 / 1M tokens for cached inputs$20.00 / 1M tokens for outputs Text:$0.60 / 1M tokens for inputs$0.06 / 1M tokens for cached inputs$2.40 / 1M tokens for outputs Image:$0.80 / 1M tokens for inputs$0.08 / 1M tokens for cached inputs - Price - $0.017 per minute / $0.00028 per second - Price - $0.0045 per minute / $0.00008 per second - Price - $0.034 per minute / $0.00057 per second - Price - $10.00 / 1k calls Search content tokens are free. - Price - Now:1 GB for $0.03 / 64GB for $1.92 per container Starting March 31, 2026:1 GB for $0.03 / 64GB for $1.92 per 20-minute session per container - Provides lower costs for requests in exchange for slower response times and occasional resource unavailability. Ideal for non-production or lower priority tasks. - Enterprise offerings - Contact our sales team to learn more about Data residency(opens in a new window), Scale Tier andReserved Capacity designed for cutting-edge customers running larger workloads. - We recommend experimenting with all of these models in the Playground(opens in a new window) to explore which models provide the best price performance trade-off for your usage. - Do you offer an enterprise package or SLAs? - We offer different tiers of access to our enterprise customers that include SLAs, lower latency, and more. Please contact our sales team to learn more. - Yes, we treat Playground usage the same as regular API usage. You will be billed at the per-token input and output prices mentioned above. - How will I know how many tokens I’ve used each month? - A token is a mathematical representation of natural language. Log in to your account to view your usage tracking dashboard(opens in a new window). This dashboard will show you how many tokens you’ve used during the current and past billing cycles. - You can set a monthly budget in your billing settings(opens in a new window), after which we’ll stop serving your requests. There may be a delay in enforcing the limit, and you are responsible for any overage incurred. You can also configure an email notification threshold to receive an email alert once you cross that threshold each month. We recommend checking your usage tracking dashboard(opens in a new window) regularly to monitor your spend. - Is access to the API included in ChatGPT Plus, Business, Enterprise or Edu? - No, OpenAI APIs are billed separately from ChatGPT Plus, Business, Enterprise and Edu. ChatGPT subscription pricing can be found at openai.com/chatgpt/pricing/. - Images are converted into tokens and charged per token. Text models price image tokens at standard text token rates, while GPT Image and gpt-realtime uses a separate image token rate. Models like gpt-4.1-mini, gpt-4.1-nano, and o4-mini convert images into tokens differently. Learn more in our docs(opens in a new window). - ChatGPT Business(opens in a new window) - ChatGPT Enterprise(opens in a new window) - ChatGPT for Education(opens in a new window) + Image 1 + Business + A secure workspace with company context and flexible seat types for any budget. + Get started(opens in a new window) + Standard seat + $20/ month + $20/month if billed annually. $25/month if billed monthly. + Premium seat + $100/ month + 5x more usage than standard, with no 5-hour limit + $100/month if billed annually. $125/month if billed monthly. + No training on your business data by default + Mix and match seat types + For teams of 2–200 employees. Unlimited subject to abuse guardrails. Learn more(opens in a new window) + Image 2 + Enterprise + Enterprise-grade AI, security, and support for businesses operating at scale
-
Anthropic model deprecations changed
Tracking since Jun 25, 2026 · 50 snapshots on file
Removed documentation of model deprecations and retirement schedules for Claude 1, Instant, and older Claude models along with API parameter deprecation information.
- Removed July 21, 2025 retirement dates for claude-2.0, claude-2.1, and claude-3-sonnet-20240229
- Removed November 6, 2024 retirement dates for Claude 1.0, 1.1, 1.2, 1.3, Instant 1.0, 1.1, and 1.2 models
- Removed deprecation status and 400 error behavior for temperature, top_p, and top_k parameters on Claude 4.7 and later models
Full history & current snapshot →View raw diff +0 −18
- On January 21, 2025, Anthropic notified developers using Claude 2, Claude 2.1, and Claude Sonnet 3 models of their upcoming retirements. - | July 21, 2025 | claude-2.0 | claude-opus-4-8 | - | July 21, 2025 | claude-2.1 | claude-opus-4-8 | - | July 21, 2025 | claude-3-sonnet-20240229 | claude-sonnet-5-5 | - 2024-09-04: Claude 1 and Instant models - On September 4, 2024, Anthropic notified developers using Claude 1 and Instant models of their upcoming retirements. - | November 6, 2024 | claude-1.0 | claude-haiku-4-5-20251001 | - | November 6, 2024 | claude-1.1 | claude-haiku-4-5-20251001 | - | November 6, 2024 | claude-1.2 | claude-haiku-4-5-20251001 | - | November 6, 2024 | claude-1.3 | claude-haiku-4-5-20251001 | - | November 6, 2024 | claude-instant-1.0 | claude-haiku-4-5-20251001 | - | November 6, 2024 | claude-instant-1.1 | claude-haiku-4-5-20251001 | - | November 6, 2024 | claude-instant-1.2 | claude-haiku-4-5-20251001 | - API parameter deprecations - Anthropic occasionally deprecates request parameters that no longer apply to current models. How the API treats a deprecated parameter depends on the model, as the following table shows. Most SDKs keep deprecated parameters in their request types so existing code continues to type-check. The Python SDK (v1.0 and later) removes temperature, top_p, and top_k, so passing them raises a TypeError. - | Parameter | Status | Behavior | Recommended replacement | - | temperature, top_p, top_k | Deprecated (Claude 4.7 and later models) | Returns a 400 error when set to a non-default value on Claude 4.7 and later models and Claude Mythos Preview. | Omit and use prompting to guide model behavior. | - For migration steps, see the migration guide.
-
NVIDIA Nemotron models changed
Tracking since Jun 25, 2026 · 40 snapshots on file
Removed Privasis-Cleaner model version 0.6B from Nemotron offerings.
- nvidia/Privasis-Cleaner-0.6B removed
- license: other
- availability date: 2026-06-08
Full history & current snapshot →View raw diff +0 −1
- nvidia/Privasis-Cleaner-0.6B license:other 2026-06-08 -
Tracking since Jun 25, 2026 · 9 snapshots on file
Gemini 3.7 Flash and 3.5 Flash models removed from batch enqueued tokens table across all tiers.
Full history & current snapshot →View raw diff +0 −6
- | Gemini 3.7 Flash | 3,000,000 | - | Gemini 3.5 Flash | 3,000,000 | - | Gemini 3.7 Flash | 400,000,000 | - | Gemini 3.5 Flash | 400,000,000 | - | Gemini 3.7 Flash | 1,000,000,000 | - | Gemini 3.5 Flash | 1,000,000,000 |
No changes match these filters.