Models

Frontier & open-weight models compared by capability — columns are flagship models, rows are the headline benchmarks that measure each one. Every number is the published figure, restated with a link to the source it came from · as of 2026-06-14. Curated matrix is 130d old — a refresh is due; live leaderboard moves still flow into Model changes below.

We cite the benchmark authorities rather than re-rank them — Epoch AI, LMArena and Artificial Analysis. ChangeRadar's job is the changes: what gets released, repriced, deprecated, or quietly shifts behind a stable model id.

Vendor champions · best flagship per vendor · GPQA Diamond

OpenAI reported

GPT-6 Astra

96GPQA Diamond

Source ↗

Google

Gemini 3.1 Pro

94.3GPQA Diamond

Compare ⇄

Moonshot AI reported

Kimi K3

93.5GPQA Diamond

Source ↗

Alibaba reported

Qwen3.8-Max

92.6GPQA Diamond

Source ↗

Anthropic reported

Claude Opus 5.5

89.9SWE-bench Pro

Source ↗

Mistral AI reported

Mistral Large 4

88.1GPQA Diamond

Source ↗

DeepSeek reported

DeepSeek-V3.2

82.4GPQA Diamond

Source ↗

Meta open

Llama 4 Maverick

69.8GPQA Diamond

Compare ⇄

xAI reported

Grok 4.7

37.6Terminal-Bench

Source ↗

Zhipu AI (Z.ai) reported

GLM-5.3

28.3Terminal-Bench

Source ↗

Each vendor's most recent publicly-available flagship, ordered by score (highest first) — Google · OpenAI · Anthropic · xAI highlighted. The weekly market-watch surfaces new releases automatically; one tagged reported is the latest release shown with vendor-reported scores (linked to source) until we independently cite it. Score = GPQA Diamond; every number links to its source.

Compare two models

Qwen3-235B-A22B-Instruct-2507 — AlibabaQwen3-235B-A22B-Thinking-2507 — AlibabaQwen3-Coder-480B-A35B-Instruct — AlibabaQwen3-Max — AlibabaQwen3.7-Max — AlibabaClaude Fable 5 — AnthropicClaude Haiku 4.5 — AnthropicClaude Mythos 5 — AnthropicClaude Opus 4.6 — AnthropicClaude Opus 4.7 — AnthropicClaude Opus 4.8 — AnthropicClaude Sonnet 4.6 — AnthropicDeepSeek-R1-0528 — DeepSeekDeepSeek-V3.1 — DeepSeekDeepSeek-V3.2 — DeepSeekDeepSeek-V3.2-Exp — DeepSeekDeepSeek-V3.2-Speciale — DeepSeekDeepSeek-V4-Pro — DeepSeekGemini 2.5 Pro — GoogleGemini 3 Flash — GoogleGemini 3 Pro — GoogleGemini 3.1 Pro — GoogleGemma 3 12B IT — GoogleGemma 3 1B IT — GoogleGemma 3 27B IT — GoogleGemma 3 4B IT — GoogleLlama 4 Behemoth — MetaLlama 4 Maverick — MetaLlama 4 Scout — MetaCodestral 25.01 — Mistral AIMagistral Medium 1.2 — Mistral AIMagistral Small 1.2 — Mistral AIMinistral 3 — Mistral AIMistral Large 3 — Mistral AIMistral Medium 3.5 — Mistral AIKimi K2 — Moonshot AIKimi K2 Thinking — Moonshot AIKimi K2.6 — Moonshot AIKimi K2.7-Code — Moonshot AIGPT-5 — OpenAIGPT-5.1 — OpenAIGPT-5.2 — OpenAIGPT-5.5 — OpenAIGrok 4 — xAIGrok 4 Fast — xAIGrok 4 Heavy — xAIGrok 4.1 — xAIGrok 4.1 Fast — xAIGrok 4.3 — xAIGLM-4.6 — Zhipu AI (Z.ai)GLM-4.6V — Zhipu AI (Z.ai)GLM-4.7 — Zhipu AI (Z.ai)GLM-5.2 — Zhipu AI (Z.ai) ⇄ Qwen3-235B-A22B-Instruct-2507 — AlibabaQwen3-235B-A22B-Thinking-2507 — AlibabaQwen3-Coder-480B-A35B-Instruct — AlibabaQwen3-Max — AlibabaQwen3.7-Max — AlibabaClaude Fable 5 — AnthropicClaude Haiku 4.5 — AnthropicClaude Mythos 5 — AnthropicClaude Opus 4.6 — AnthropicClaude Opus 4.7 — AnthropicClaude Opus 4.8 — AnthropicClaude Sonnet 4.6 — AnthropicDeepSeek-R1-0528 — DeepSeekDeepSeek-V3.1 — DeepSeekDeepSeek-V3.2 — DeepSeekDeepSeek-V3.2-Exp — DeepSeekDeepSeek-V3.2-Speciale — DeepSeekDeepSeek-V4-Pro — DeepSeekGemini 2.5 Pro — GoogleGemini 3 Flash — GoogleGemini 3 Pro — GoogleGemini 3.1 Pro — GoogleGemma 3 12B IT — GoogleGemma 3 1B IT — GoogleGemma 3 27B IT — GoogleGemma 3 4B IT — GoogleLlama 4 Behemoth — MetaLlama 4 Maverick — MetaLlama 4 Scout — MetaCodestral 25.01 — Mistral AIMagistral Medium 1.2 — Mistral AIMagistral Small 1.2 — Mistral AIMinistral 3 — Mistral AIMistral Large 3 — Mistral AIMistral Medium 3.5 — Mistral AIKimi K2 — Moonshot AIKimi K2 Thinking — Moonshot AIKimi K2.6 — Moonshot AIKimi K2.7-Code — Moonshot AIGPT-5 — OpenAIGPT-5.1 — OpenAIGPT-5.2 — OpenAIGPT-5.5 — OpenAIGrok 4 — xAIGrok 4 Fast — xAIGrok 4 Heavy — xAIGrok 4.1 — xAIGrok 4.1 Fast — xAIGrok 4.3 — xAIGLM-4.6 — Zhipu AI (Z.ai)GLM-4.6V — Zhipu AI (Z.ai)GLM-4.7 — Zhipu AI (Z.ai)GLM-5.2 — Zhipu AI (Z.ai)

2 wins · Claude Fable 5 1 wins · Gemini 3.1 Pro 0 ties

Agentic coding

Benchmark	Claude Fable 5	Gemini 3.1 Pro	Δ
SWE-bench Verified i% resolved (pass@1)	95	80.6	+14.4
SWE-bench Pro i% resolved (pass@1)	80.3	54.2	+26.1

3 coverage gaps — only one model reports these

• Terminal-Bench — Gemini 3.1 Pro 68.5
• LiveCodeBench — Gemini 3.1 Pro 2887
• FrontierCode — Claude Fable 5 29.3

Tool use & agents

5 coverage gaps — only one model reports these

• TAU-bench — Gemini 3.1 Pro 99.3
• OSWorld — Claude Fable 5 85
• BrowseComp — Gemini 3.1 Pro 85.9
• GDPval-AA — Claude Fable 5 1932
• MCP Atlas — Gemini 3.1 Pro 69.2

Science & reasoning

Benchmark	Claude Fable 5	Gemini 3.1 Pro	Δ
GPQA Diamond i% accuracy	92.6	94.3	-1.7

2 coverage gaps — only one model reports these

• Humanity's Last Exam — Gemini 3.1 Pro 44.4
• ARC-AGI-2 — Gemini 3.1 Pro 77.1

General knowledge

1 coverage gap — only one model reports these

• MMMLU — Gemini 3.1 Pro 92.6

Multimodal

2 coverage gaps — only one model reports these

• MMMU — Gemini 3.1 Pro 80.5
• GDP.pdf — Claude Fable 5 29.8

Long context

1 coverage gap — only one model reports these

• MRCR — Gemini 3.1 Pro 84.9

Shared benchmarks first (with Δ when both report the same scale); one-sided coverage collapses below. Pick any two models — or a champion above — and the URL becomes shareable.

All models · benchmark matrix

⌕

Vendor

AnthropicOpenAIGoogleMoonshot AIMetaxAIDeepSeekAlibabaZhipu AI (Z.ai)Mistral AI

Capability

Open weightsAgentic codingTool use & agentsMathScience & reasoningGeneral knowledgeMultimodalLong contextHuman preference

21 of 21 columns shown

reported columns (Claude Opus 5.5, GPT-6 Astra, Grok 4.7, Qwen3.8-Max, DeepSeek-V3.2, Mistral Large 4, Kimi K3, GLM-5.3) are auto-discovered by our weekly market-watch from each vendor's own reported numbers — not independently verified, and shown when a vendor ships a model newer than the hand-cited column beside it. Full claim sets are in Vendor-reported benchmarks below.

Agentic coding

Can the model fix real bugs, ship features, and operate a dev environment end-to-end as a coding agent — the single most-watched capability in 2026 vendor launches.

Benchmark	Fable 5	Opus 5.5 reported	Opus 4.8	Mythos 5	GPT-5.5	GPT-6 Astra reported	Gemini 3.1 Pro	Grok 4.3	Grok 4.7 reported	DeepSeek-V4-Pro open	DeepSeek-V3.2 reported	Qwen3.7-Max	Qwen3.8-Max reported	Kimi K2.6 open	Kimi K3 reported	GLM-5.2 open	GLM-5.3 reported	Llama 4 Maverick open	Mistral Medium 3.5 open	Mistral Large 4 reported	Gemma 3 27B open
SWE-bench Verified i % resolved (pass@1)	95	—	88.6	—	—	—	80.6	—	—	80.6	73.1	80.4	—	80.2	76.8	—	—	—	77.6	—	—
SWE-bench Pro i % resolved (pass@1)	80.3	89.9	69.2	77.8	58.6	—	54.2	—	—	—	—	60.6	67.7	58.6	—	62.1	—	—	—	—	—
Terminal-Bench i % solved (pass@1)	—	66.4	74.6	88	82.7	57.9	68.5	—	37.6	67.9	46.4	69.7	86.6	66.7	88.3	81	28.3	—	—	28.3	—
LiveCodeBench i % pass@1	—	—	—	—	—	—	2887	—	—	93.5	83.3	—	—	89.6	—	—	—	43.4	—	—	29.7

Tool use & agents

Beyond writing code: can the model select and chain tools, follow policy, drive a computer/browser, and complete long-horizon multi-step tasks.

Benchmark	Fable 5	Opus 5.5 reported	Opus 4.8	Mythos 5	GPT-5.5	GPT-6 Astra reported	Gemini 3.1 Pro	Grok 4.3	Grok 4.7 reported	DeepSeek-V4-Pro open	DeepSeek-V3.2 reported	Qwen3.7-Max	Qwen3.8-Max reported	Kimi K2.6 open	Kimi K3 reported	GLM-5.2 open	GLM-5.3 reported	Llama 4 Maverick open	Mistral Medium 3.5 open	Mistral Large 4 reported	Gemma 3 27B open
TAU-bench i % pass / pass^k	—	—	—	—	—	—	99.3	98	—	—	—	—	—	—	—	—	—	—	—	—	—
OSWorld i % success	85	81.8	83.4	—	78.7	72.6	—	—	—	—	—	—	86.1	73.1	—	—	—	—	—	—	—
BrowseComp i % accuracy	—	—	84.3	—	90.1	—	85.9	—	—	—	—	—	—	83.2	91.2	—	—	—	—	—	—

Math

Competition and research-level mathematical reasoning, increasingly reported on uncontaminated/post-cutoff problem sets.

Benchmark	Fable 5	Opus 5.5 reported	Opus 4.8	Mythos 5	GPT-5.5	GPT-6 Astra reported	Gemini 3.1 Pro	Grok 4.3	Grok 4.7 reported	DeepSeek-V4-Pro open	DeepSeek-V3.2 reported	Qwen3.7-Max	Qwen3.8-Max reported	Kimi K2.6 open	Kimi K3 reported	GLM-5.2 open	GLM-5.3 reported	Llama 4 Maverick open	Mistral Medium 3.5 open	Mistral Large 4 reported	Gemma 3 27B open
HMMT i % accuracy (pass@1)	—	—	—	—	—	—	—	—	—	—	92.5	97.1	—	92.7	—	—	—	—	—	—	—
MATH i % accuracy	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	61.2	—	—	89

Science & reasoning

Expert-level, Google-proof reasoning across the sciences and broad academia — the benchmarks vendors point to when claiming 'PhD-level' or 'frontier' reasoning.

Benchmark	Fable 5	Opus 5.5 reported	Opus 4.8	Mythos 5	GPT-5.5	GPT-6 Astra reported	Gemini 3.1 Pro	Grok 4.3	Grok 4.7 reported	DeepSeek-V4-Pro open	DeepSeek-V3.2 reported	Qwen3.7-Max	Qwen3.8-Max reported	Kimi K2.6 open	Kimi K3 reported	GLM-5.2 open	GLM-5.3 reported	Llama 4 Maverick open	Mistral Medium 3.5 open	Mistral Large 4 reported	Gemma 3 27B open
GPQA Diamond i % accuracy	92.6	—	93.6	—	93.6	96	94.3	90.1	—	90.1	82.4	92.4	92.6	90.5	93.5	91.2	—	69.8	—	88.1	42.4
Humanity's Last Exam i % accuracy	—	67.7	57.9	64.5	57.2	57.2	44.4	—	—	37.7	25.1	41.4	43.6	54	—	40.5	—	—	—	35	—
ARC-AGI-2 i % solved	—	—	—	—	85	—	77.1	—	—	—	—	—	—	—	94.5	—	—	—	—	—	—

General knowledge

Broad multi-subject factual and reasoning coverage; the classic 'how much does it know' bucket, now reported via the harder Pro variant since base MMLU is saturated.

Benchmark	Fable 5	Opus 5.5 reported	Opus 4.8	Mythos 5	GPT-5.5	GPT-6 Astra reported	Gemini 3.1 Pro	Grok 4.3	Grok 4.7 reported	DeepSeek-V4-Pro open	DeepSeek-V3.2 reported	Qwen3.7-Max	Qwen3.8-Max reported	Kimi K2.6 open	Kimi K3 reported	GLM-5.2 open	GLM-5.3 reported	Llama 4 Maverick open	Mistral Medium 3.5 open	Mistral Large 4 reported	Gemma 3 27B open
MMLU-Pro i % accuracy	—	—	—	—	—	—	—	—	—	87.5	85	—	—	—	—	—	—	80.5	—	—	67.5

Multimodal

Vision + language understanding and visual reasoning — how well the model interprets images, diagrams, charts and figures.

Benchmark	Fable 5	Opus 5.5 reported	Opus 4.8	Mythos 5	GPT-5.5	GPT-6 Astra reported	Gemini 3.1 Pro	Grok 4.3	Grok 4.7 reported	DeepSeek-V4-Pro open	DeepSeek-V3.2 reported	Qwen3.7-Max	Qwen3.8-Max reported	Kimi K2.6 open	Kimi K3 reported	GLM-5.2 open	GLM-5.3 reported	Llama 4 Maverick open	Mistral Medium 3.5 open	Mistral Large 4 reported	Gemma 3 27B open
MMMU i % accuracy	—	—	—	—	83.2	—	80.5	—	—	—	—	—	—	79.4	—	—	—	73.4	—	—	64.9

Long context

Retrieval and reasoning quality as context length grows into the hundreds-of-thousands / millions of tokens — beyond simple needle-in-a-haystack.

Benchmark	Fable 5	Opus 5.5 reported	Opus 4.8	Mythos 5	GPT-5.5	GPT-6 Astra reported	Gemini 3.1 Pro	Grok 4.3	Grok 4.7 reported	DeepSeek-V4-Pro open	DeepSeek-V3.2 reported	Qwen3.7-Max	Qwen3.8-Max reported	Kimi K2.6 open	Kimi K3 reported	GLM-5.2 open	GLM-5.3 reported	Llama 4 Maverick open	Mistral Medium 3.5 open	Mistral Large 4 reported	Gemma 3 27B open
MRCR i % accuracy	—	—	—	—	74	—	84.9	—	—	—	—	—	—	—	—	—	—	—	—	—	—

Human preference

Aggregate real-user preference from blind head-to-head comparisons — the closest thing to a 'do people actually like the answers' metric, and the one number vendors love to top.

Benchmark	Fable 5	Opus 5.5 reported	Opus 4.8	Mythos 5	GPT-5.5	GPT-6 Astra reported	Gemini 3.1 Pro	Grok 4.3	Grok 4.7 reported	DeepSeek-V4-Pro open	DeepSeek-V3.2 reported	Qwen3.7-Max	Qwen3.8-Max reported	Kimi K2.6 open	Kimi K3 reported	GLM-5.2 open	GLM-5.3 reported	Llama 4 Maverick open	Mistral Medium 3.5 open	Mistral Large 4 reported	Gemma 3 27B open
LMArena Elo i Elo	—	—	—	1458	1474	—	—	—	—	—	—	—	—	1466	—	—	—	1417	—	—	1338

Showing 13 flagship models across the headline benchmarks; 53 models and 84 benchmarks tracked in total from 60 primary & aggregator sources. Claude Mythos 5 numbers are limited (access-restricted preview). Numbers are published facts, restated with a per-cell source link; vendor benchmark charts are linked to their source, not rehosted.
Sources include: 9to5Google (Gemini 3 Flash launch coverage) · AIFire (citing OpenAI GPT-5.2 release) · Artificial Analysis · BenchLM.ai (citing LMArena) · BinaryVerse AI (xAI official figures) · BuildFastWithAI (citing OpenAI GPT-5.5 launch) · Caylent · Codersera (reporting Moonshot's figures) · DataCamp · DataCamp (reproducing Meta's Llama 4 launch chart) · DeepSeek-AI (arXiv 2512.02556) · DeepSeek-AI (Hugging Face model card) · Google (official Gemini 2.5 launch blog) · Google (official Gemini 3 Flash launch blog) · Google (official Gemini 3 launch blog) · Google DeepMind (Gemini 3.1 Pro model card) · +36 more.

Vendor-reported benchmarks

Numbers as claimed by the vendor on their own model/system card — not independently verified and often measured with a favourable harness. We track each vendor's claims over time and link to the source; cross-check against the cited matrix above.

Claude Opus 5.5 Anthropic

SWE-bench Pro	89.9%
Terminal-Bench 4.0 · xhigh effort; standard error ±2.6 pts	66.4%
FrontierCode v1.1 Main	54.4%
CursorBench 4.0	57.8%
GDPval-AA v2.1	1846 Elo
Humanity's Last Exam · with tools	67.7%
Terminal-Bench-Science 0.1	58.7%
OSWorld 2.1 · partial credit	81.8%
Chartography · with tools	89.0%
AutomationBench	40.0%
SWE-bench Multilingual	93.9%
ArXivMath · without tools; 96.9% with tools	91.2%

vendor card ↗

GPT-6 Astra OpenAI

GPQA Diamond	96.0%
FrontierMath Tier 4 (v2)	97.6%
ARC-AGI-3 · adapter harness	99.9%
Terminal-Bench 4.0	57.9%
Terminal-Bench Science 0.1	64.6%
DeepSWE v1.1	74.1%
Humanity's Last Exam (w/ tools)	57.2%
OSWorld 2.0 (partial)	72.6%
Agents' Last Exam	59.3%
ExploitBench	100.0%

vendor card ↗

Grok 4.7 xAI

CursorBench 4.0 · xhigh reasoning effort	46.3%
DeepSWE v1.1 · high effort	71.0%
Terminal-Bench 4.0 · xhigh reasoning effort	37.6%
EEBench · electrical engineering	64.0%
AA-Briefcase v1.1 · xhigh reasoning effort	1657 Elo
Harvey Legal Agent Benchmark · legal tasks	19.6%
HealthBench Professional · clinical reasoning	56.7%

vendor card ↗

Gemini 3.1 Pro Google

GPQA Diamond · Single attempt (pass@1), graduate-level science	94.3%
ARC-AGI-2 · Verified by ARC Prize, abstract reasoning	77.1%
Humanity's Last Exam · Without tools	44.4%
SWE-Bench Verified · Software engineering tasks	80.6%
LiveCodeBench Pro · Competitive programming	2887 Elo
Terminal-Bench 2.0 · Standard harness, agentic coding	68.5%
ARC-AGI-2 · Verified abstract reasoning test	77.1%
MRCR v2 128k · Long context understanding	84.9%
APEX-Agents · Agentic task performance	33.5%
τ-Bench Telecom · Telecom domain benchmark	99.3%

vendor card ↗

Qwen3.8-Max Alibaba

Terminal Bench 2.1 · Agentic terminal task completion	86.6 Elo
SWE-bench Pro · Enterprise-scale software engineering	67.7 pass@1
GPQA Diamond · Graduate-level science QA	92.6 %
PaperBench · Research paper reproduction and understanding	93.0 %
IFBench · Instruction following and constraints	82.8 %
OSWorld-Verified · Desktop/GUI computer use operations	86.1 %
FrontierSWE · Extended engineering project work	73.5 %
Humanity's Last Exam (HLE) · Multidisciplinary expert-level questions	43.6 %
OmniDocBench 1.5 · Document understanding with complex layouts	92.1 %
Parametric CAD Bench · Structured CAD task reasoning	91.5 %
DeepSWE 1.1 · Agentic coding harness (Claude Code tools)	56.6 %

vendor card ↗

DeepSeek-V3.2 DeepSeek

MMLU-Pro · Thinking mode	85.0 %
GPQA-Diamond · Thinking mode	82.4 %
Humanity's Last Exam · Thinking mode, text-only	25.1 %
LiveCodeBench · COT, data from 2024.08-2025.04	83.3 pass@1
Codeforces · Thinking mode	2386 Rating
AIME 2025 · Thinking mode	93.1 %
HMMT Feb 2025 · Thinking mode	92.5 %
Terminal Bench 2.0 · Code Agent, Thinking mode	46.4 %
SWE-bench Verified · Code Agent, Thinking mode	73.1 Resolved
SWE-bench Multilingual · Code Agent, Thinking mode	70.2 Resolved

vendor card ↗

Llama 4 Maverick Meta

MMLU · instruction-tuned	85.5%
MMLU Pro · instruction-tuned	80.5%
GPQA Diamond · instruction-tuned	69.8%
LiveCodeBench · instruction-tuned	43.4%
AIME 2025 · instruction-tuned	39.0%
MMMU · image reasoning; instruction-tuned	73.4%
MathVista · image reasoning; instruction-tuned	73.7%
ChartQA · image understanding; instruction-tuned	90%
DocVQA · image understanding test split; instruction-tuned	94.4%
LMArena Elo · experimental chat variant; vendor-claimed	1417 Elo
Terminal-Bench 2.1 · agentic task success	7.9%
Terminal-Bench Hard · agentic task success	6.8%

vendor card ↗

Mistral Large 4 Mistral AI

DeepSWE v1.1 · Artificial Analysis evaluation	61.7%
SWE-Atlas-QnA · Artificial Analysis evaluation	59.4%
Terminal-Bench 4 · Artificial Analysis evaluation	28.3%
Coding Agent Index · Average of DeepSWE, SWE-Atlas, Terminal-Bench	49.8%
AutomationBench · 657 business workflows	59.9%
AA-Briefcase · Long-horizon knowledge work	1393 Elo
Cybench · 40 security competition challenges	93%
Artificial Analysis Cyber Index - Reproduce-and-patch · Reproduce vulnerability and patch it	82%
Lakera B3 AI Security Benchmark · Attack resistance rate	93.3%
GPQA Diamond · From Epoch AI Benchmarking Hub	88.1%
LiveBench 2.0 Reasoning	83.9%
Humanity's Last Exam · Artificial Analysis measurement	35.0%

vendor card ↗

Kimi K3 Moonshot AI

GPQA Diamond · max reasoning effort	93.5%
AIME 2025 · max reasoning effort	96.1%
SWE-Bench Verified · max reasoning effort	76.8%
Terminal-Bench 2.1 · max reasoning effort	88.3%
FrontierSWE · max reasoning effort	81.2%
DeepSWE · Kimi Code configuration	67.5%
SWE-Marathon · max reasoning effort	42.0%
Program Bench · max reasoning effort	77.8%
BrowseComp · agentic configuration	91.2%
GDPval-AA v2 · max reasoning effort	1687 Elo
HLE-Full · no tools, max reasoning effort	43.5%
ARC-AGI · ARC-AGI semi-private evaluation	94.5%

vendor card ↗

GLM-5.3 Zhipu AI (Z.ai)

Terminal-Bench 3.0	28.3 score
DeepSWE v1.1	66.9 %
Terminal-Bench 2.1	88.2 score
CyberGym	84.5%
Agents' Last Exam (CLI)	28.5 score
ExploitBench	54.4%

vendor card ↗

Model changes

New releases, deprecations, and benchmark score moves we’ve recorded — newest first.

• Benchmark LMArena (text) Oct 23, 2026

  LMArena (text) scores changed

  Arena Elo (text, overall) updated.

• Model MiniMax Oct 23, 2026

  MiniMax models changed

  MiniMax-M3.1-Flash model added with license:other and availability date of 2026-10-17

• Model NVIDIA Nemotron Oct 23, 2026

  NVIDIA Nemotron models changed

  nvidia/Flora-Captioner license type changed from unknown to other.

• Vendor claim Mistral — reported benchmarks Oct 22, 2026

  Mistral AI reported benchmarks updated

  Mistral Large 4: 12 benchmark claims (via web search)

• Benchmark GPQA Diamond (Epoch) Oct 22, 2026

  GPQA Diamond (Epoch) scores changed

  GPQA Diamond accuracy (0–1) updated.

• Model AI2 OLMo Oct 22, 2026

  AI2 OLMo models changed

  Removed model allenai/intent-aware-lfqa-qwen3-4b-multiview with odc-by license and expiration date 2026-04-13.

• Benchmark LMArena (text) Oct 21, 2026

  LMArena (text) scores changed

  Arena Elo (text, overall) updated.

• Model NVIDIA Nemotron Oct 21, 2026

  NVIDIA Nemotron models changed

  Removed MiniMax-M3-DSpark model with 'other' license (expiring 2026-07-15) from NVIDIA Nemotron offerings.

• Model Microsoft Phi Oct 21, 2026

  Microsoft Phi models changed

  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 NVIDIA Nemotron Oct 20, 2026

  NVIDIA Nemotron models changed

  Removed two NVIDIA models and added Flora-Captioner model with unknown license status, with effective date of 2026-10-01.

• Model Microsoft Phi Oct 20, 2026

  Microsoft Phi models changed

  Removed microsoft/Dayhoff-3b-GR-HM-41000 model with MIT license from available models.

• Vendor claim Meta Llama — reported benchmarks Oct 20, 2026

  Meta reported benchmarks updated

  Llama 4 Maverick: 12 benchmark claims (via web search)

• Vendor claim Moonshot Kimi — reported benchmarks Oct 19, 2026

  Moonshot AI reported benchmarks updated

  Kimi K3: 12 benchmark claims (via web search)

• Vendor claim Zhipu GLM — reported benchmarks Oct 19, 2026

  Zhipu AI (Z.ai) reported benchmarks updated

  GLM-5.3: 6 benchmark claims (via web search)

• Vendor claim Qwen — reported benchmarks Oct 19, 2026

  Alibaba reported benchmarks updated

  Qwen3.8-Max: 11 benchmark claims (via web search)

• Vendor claim DeepSeek — reported benchmarks Oct 19, 2026

  DeepSeek reported benchmarks updated

  DeepSeek-V3.2: 10 benchmark claims (via web search)

• Vendor claim xAI — reported benchmarks Oct 19, 2026

  xAI reported benchmarks updated

  Grok 4.7: 7 benchmark claims (via web search)

• Vendor claim Google — reported benchmarks Oct 19, 2026

  Google reported benchmarks updated

  Gemini 3.1 Pro: 10 benchmark claims (via web search)

• Vendor claim OpenAI — reported benchmarks Oct 19, 2026

  OpenAI reported benchmarks updated

  GPT-6 Astra: 10 benchmark claims (via web search)

• Vendor claim Anthropic — reported benchmarks Oct 19, 2026

  Anthropic reported benchmarks updated

  Claude Opus 5.5: 12 benchmark claims (via web search)

• Model NVIDIA Nemotron Oct 19, 2026

  NVIDIA Nemotron models changed

  Removed 6 NVIDIA Nemotron and SDLLM model entries from the catalog.

• Model NVIDIA Nemotron Oct 18, 2026

  NVIDIA Nemotron models changed

  Added Nemotron-3.5-Parse model with openmdw-1.1 license and expiration date of 2026-09-03.

• Model DeepSeek (weights) Oct 17, 2026

  DeepSeek (weights) models changed

  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 NVIDIA Nemotron Oct 17, 2026

  NVIDIA Nemotron models changed

  New model variant added: nvidia/NVIDIA-Nemotron-Labs-Dory1.0-4B-BF16 with other license and availability date 2026-10-16.

• Model NVIDIA Nemotron Oct 16, 2026

  NVIDIA Nemotron models changed

  NVIDIA Nemotron Parse 2.0 model listing was removed.