> Where every number on The Model Index comes from: the sources, how the index score is calculated, its known limits, and the full release list.

Every number, sourced

# Data & method

Where each number comes from, how the index is built, and how the release list stays current.

240

major releases tracked

203

linked to an announcement

141

dates checked on the lab's own page

53/69

lab pages read, 1 Oct 2026

## How the index works · v1.0

Evidence. Only independent results: the Epoch AI benchmarking hub, which runs evaluations itself or collects them from independent leaderboards. Results a lab reported about its own model (technical reports, system cards) are removed: 564 of them.

Core tests. Six areas, shown on every profile: reasoning & knowledge (GPQA Diamond, Humanity's Last Exam), maths (FrontierMath), coding (SWE-bench Verified, Terminal-Bench), agents (APEX-Agents), novel problems (ARC-AGI-2) and long tasks (METR time horizon). About forty further independent tests also inform the score, so older and less-tested models can still be placed.

Model. An item response model estimates one general ability per model from every test it has taken, allowing for how hard and how discriminating each test is. It is a single latent score, not an average of the six areas. GPT-4 (March 2023) = 100; 50 points is one unit of ability.

Uncertainty. Each score has a 90% band from the fit, measured against GPT-4. It assumes the test parameters are right, so treat it as a lower bound on the real uncertainty. Rated: core results in three areas (or core results plus anchors across three). Provisional: evidence in two areas.

Estimates. Models Epoch has not tested yet get an estimate from their Artificial Analysis score, calibrated on 25 models that have both (R² 0.899). Estimates carry a 90% prediction interval, are marked ~, and are only made inside the calibrated range (AA 11–58).

World affordability. On the World page, a $20-a-month plan is set against the median household living standard per person from the latest World Bank household survey for each country, revalued from 2021 purchasing-power dollars to recent US dollars (PPP × local inflation since 2021 ÷ official exchange rate), assuming real living standards unchanged since the survey. "Over a tenth" is the share of people below $2,400 a year on the same basis. Surveys older than 2015, and countries whose official exchange rate is not what people can get, are left out. The 10% line is our threshold, not a test of who can pay. The method was reviewed by a second model (GPT-6 Astra) and its corrections applied.

Cross-check. Rankings agree almost perfectly with Epoch's own Capabilities Index, which uses the same evidence. The value here is transparency: every score shows the tests behind it, broken down by area, alongside release dates, prices and sources.

## Sources

Capability results

[Epoch AI benchmarking hub ↗](https://epoch.ai/benchmarks)

Second opinion

[Artificial Analysis ↗](https://artificialanalysis.ai/leaderboards/models)

Prices, context

[OpenRouter model list ↗](https://openrouter.ai/models)

Task length

[METR time horizons ↗](https://metr.org/time-horizons)

Releases

Labs' own announcements

Safety findings

UK AISI, US CAISI, METR, lab system cards

A monitor reads 69 official lab pages every 30 minutes and checks the public leaderboards for anything missed. Clear launches on a lab's own site are added automatically; anything uncertain waits for review.

## Every release · 240 rows

| DATE | LAB | RELEASE | TYPE | SOURCE |
| --- | --- | --- | --- | --- |
| 30 Sep 2026 | Google | Gemini 4 ArgonAUTO | flagship | [blog.google ↗](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-4-argon/) |
| 29 Sep 2026 | OpenAI | GPT-6.1 Sol | flagship | [openai.com ↗](https://openai.com/index/introducing-gpt-6-1-sol/) |
| 28 Sep 2026 | Anthropic | Claude Sonnet 5.5 | flagship | [anthropic.com ✓ ↗](https://www.anthropic.com/claude-sonnet-5-5) |
| 22 Sep 2026 | OpenAI | GPT-6 Sol + GPT-6 Luna | flagship | [openai.com ↗](https://openai.com/index/introducing-gpt-6-sol-and-luna/) |
| 22 Sep 2026 | Anthropic | Claude Opus 5.5 | flagship | [anthropic.com ✓ ↗](https://www.anthropic.com/claude-opus-5-5) |
| 21 Sep 2026 | xAI | Grok 4.7 | flagship | [x.ai ✓ ↗](https://x.ai/news/grok-4-7) |
| 21 Sep 2026 | Xiaomi | MiMo-V2.6-Pro | open-weights | [huggingface.co ✓ ↗](https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Pro-RL) |
| 21 Sep 2026 | Xiaomi | MiMo-V2.6-Flash | open-weights | [huggingface.co ✓ ↗](https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Flash-RL) |
| 18 Sep 2026 | StepFun | Step 5 Preview | flagship | [platform.stepfun.ai ↗](https://platform.stepfun.ai/docs/en/guides/models/step-5-preview) |
| 11 Sep 2026 | Moonshot | Kimi K2.8 Preview | flagship | – |
| 10 Sep 2026 | DeepSeek | DeepSeek-V4.1-Flash | open-weights | [api-docs.deepseek.com ↗](https://api-docs.deepseek.com/news/news260910/) |
| 4 Sep 2026 | OpenAI | GPT-6 Astra (preview 09-03) | flagship | [openai.com ↗](https://openai.com/index/gpt-6-astra-next-generation-work/) |
| 2 Sep 2026 | Meta | Muse Spark 1.3 | flagship | – |
| 2 Sep 2026 | Google | Gemini 3.8 Flash | flagship | [blog.google ✓ ↗](https://blog.google/innovation-and-ai/models-and-research/gemini-models/3-8-flash-and-3-8-flash-cyber/) |
| 1 Sep 2026 | Anthropic | Claude Fable 5.1 + Mythos 5.1 | flagship | [anthropic.com ↗](https://www.anthropic.com/claude-fable-and-mythos-5-1) |
| 28 Aug 2026 | Tencent | Hy4 preview | open-weights | [simonwillison.net ↗](https://simonwillison.net/2026/Aug/29/hy4/) |
| 26 Aug 2026 | Alibaba | Qwen3.8-Flash | small | [qwenlm.github.io ↗](https://qwenlm.github.io/qwen-code-docs/en/blog/updates/weekly-update-2026-09-03/) |
| 24 Aug 2026 | Alibaba | Qwen3.8-Flash-Next | open-weights | [huggingface.co ✓ ↗](https://huggingface.co/Qwen/Qwen3.8-Flash-Next) |
| 21 Aug 2026 | DeepSeek | DeepSeek-V4-Flash-Vision-Exp (weights 08-31) | open-weights | [huggingface.co ↗](https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-Vision-Exp) |
| 20 Aug 2026 | Zhipu | GLM-5.3-Flash | open-weights | [huggingface.co ↗](https://huggingface.co/zai-org/GLM-5.3-Flash) |
| 14 Aug 2026 | Zhipu | GLM-5.3 | open-weights | [z.ai ✓ ↗](https://z.ai/blog/glm-5.3) |
| 14 Aug 2026 | Alibaba | Qwen3.8-27B | open-weights | [qwen.ai ↗](https://qwen.ai/blog?id=qwen3.8-flash-next&ref=upstract.com) |
| 13 Aug 2026 | Google | Gemini 3.7 Flash | flagship | [blog.google ✓ ↗](https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-3-7-flash/) |
| 13 Aug 2026 | DeepSeek | DeepSeek-V4-Pro (GA) | flagship | [api-docs.deepseek.com ↗](https://api-docs.deepseek.com/updates/) |
| 12 Aug 2026 | xAI | Grok 4.6 | flagship | [x.ai ✓ ↗](https://x.ai/news/grok-4-6) |
| 12 Aug 2026 | ByteDance | Seed 2.1 | flagship | – |
| 12 Aug 2026 | Alibaba | Qwen3.8-2.4T-A95B | open-weights | [huggingface.co ↗](https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B) |
| 10 Aug 2026 | Meta | Muse Glimmer 30B | open-weights | – |
| 5 Aug 2026 | Meta | Muse Spark 1.2 (+ Muse Code) | flagship | – |
| 3 Aug 2026 | Alibaba | Qwen3.8-Max (preview 07-19; weights 08-12) | flagship | [qwenlm.github.io ↗](https://qwenlm.github.io/qwen-code-docs/en/blog/cases/qwencode-bailian-skill-openai-cover-gen/) |
| 31 Jul 2026 | DeepSeek | DeepSeek-V4-Flash (official release; 0731 weights) | open-weights | [api-docs.deepseek.com ✓ ↗](https://api-docs.deepseek.com/updates/) |
| 30 Jul 2026 | TM Thinking Machines | Inkling-Small | open-weights | [thinkingmachines.ai ✓ ↗](https://thinkingmachines.ai/news/inkling-small/) |
| 24 Jul 2026 | Anthropic | Claude Opus 5 | flagship | [anthropic.com ✓ ↗](https://www.anthropic.com/news/claude-opus-5) |
| 23 Jul 2026 | Ant Group | Ling 3.0 | open-weights | – |
| 21 Jul 2026 | Google | Gemini 3.6 Flash + 3.5 Flash-Lite | flagship | [blog.google ✓ ↗](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-6-flash-3-5-flash-lite-3-5-flash-cyber/) |
| 20 Jul 2026 | Meituan | LongCat 2.0 | open-weights | – |
| 16 Jul 2026 | Moonshot | Kimi K3 (weights 07-27) | open-weights | [kimi.com ✓ ↗](https://www.kimi.com/blog/kimi-k3?ref=apifox.com) |
| 15 Jul 2026 | TM Thinking Machines | Inkling | open-weights | [thinkingmachines.ai ✓ ↗](https://thinkingmachines.ai/news/introducing-inkling/) |
| 9 Jul 2026 | OpenAI | GPT-5.6 (Sol/Terra/Luna; preview 06-26) | flagship | [openai.com ↗](https://openai.com/index/improving-gpt-5-6-sol-in-chatgpt/) |
| 9 Jul 2026 | Meta | Muse Spark 1.1 (Meta Model API preview) | flagship | [about.fb.com ↗](https://about.fb.com/news/2026/07/meta-ai-muse-spark-doesnt-just-think-it-acts/) |
| 8 Jul 2026 | xAI | Grok 4.5 | flagship | [x.ai ↗](https://x.ai/news/grok-4-5) |
| 30 Jun 2026 | Anthropic | Claude Sonnet 5 | flagship | [anthropic.com ✓ ↗](https://www.anthropic.com/news/claude-sonnet-5) |
| 16 Jun 2026 | Zhipu | GLM-5.2 | open-weights | [z.ai ✓ ↗](https://z.ai/blog/glm-5.2) |
| 12 Jun 2026 | Moonshot | Kimi K2.7 Code | open-weights | [kimi.com ✓ ↗](https://www.kimi.com/resources/kimi-k2-7-code) |
| 9 Jun 2026 | Anthropic | Claude Fable 5 + Mythos 5 (restricted) | flagship | [anthropic.com ✓ ↗](https://www.anthropic.com/news/claude-fable-5-mythos-5) |
| 4 Jun 2026 | NVIDIA | Nemotron 3 Ultra | open-weights | [developer.nvidia.com ✓ ↗](https://developer.nvidia.com/blog/nvidia-nemotron-3-ultra-powers-faster-more-efficient-reasoning-for-long-running-agents/) |
| 1 Jun 2026 | MiniMax | MiniMax-M3 | open-weights | [huggingface.co ✓ ↗](https://huggingface.co/MiniMaxAI/MiniMax-M3) |
| 1 Jun 2026 | Alibaba | Qwen3.7-Plus | flagship | – |
| 29 May 2026 | StepFun | Step 3.7 Flash | open-weights | [huggingface.co ↗](https://huggingface.co/stepfun-ai/Step-3.7-Flash) |
| 28 May 2026 | Anthropic | Claude Opus 4.8 | flagship | [anthropic.com ✓ ↗](https://www.anthropic.com/research/claude-opus-4-8) |
| 20 May 2026 | xAI | Grok Build 0.1 (coding) | reasoning | [x.ai ↗](https://x.ai/news/grok-build-0-1) |
| 19 May 2026 | Google | Gemini 3.5 Flash (3.5 Pro announced; unreleased) | flagship | [blog.google ✓ ↗](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-5/?_bhlid=25af37ed31cbc0806de888e382fe92e66a1318da&trk=article-ssr-frontend-pulse_little-text-block) |
| 19 May 2026 | Alibaba | Qwen3.7-Max | flagship | – |
| 28 Apr 2026 | Mistral | Mistral Medium 3.5 | open-weights | [mistral.ai ↗](https://mistral.ai/news/vibe-remote-agents-mistral-medium-3-5/) |
| 24 Apr 2026 | DeepSeek | DeepSeek-V4 preview (V4-Pro/V4-Flash) | open-weights | – |
| 23 Apr 2026 | Tencent | Hy3 preview | open-weights | [huggingface.co ✓ ↗](https://huggingface.co/tencent/Hy3-preview) |
| 23 Apr 2026 | OpenAI | GPT-5.5 (Thinking/Pro) | flagship | [openai.com ✓ ↗](https://openai.com/index/introducing-gpt-5-5/) |
| 22 Apr 2026 | Xiaomi | MiMo-V2.5-Pro | open-weights | [huggingface.co ↗](https://huggingface.co/XiaomiMiMo/MiMo-V2.5-Pro) |
| 20 Apr 2026 | Moonshot | Kimi K2.6 | open-weights | [kimi.com ✓ ↗](https://www.kimi.com/en/blog/kimi-k2-6) |
| 20 Apr 2026 | Alibaba | Qwen3.6-Max-Preview | flagship | – |
| 17 Apr 2026 | xAI | Grok 4.3 (beta; API/GA 04-30) | flagship | [docs.x.ai ↗](https://docs.x.ai/developers/models/grok-4.3) |
| 16 Apr 2026 | Anthropic | Claude Opus 4.7 | flagship | [anthropic.com ✓ ↗](https://www.anthropic.com/news/claude-opus-4-7) |
| 16 Apr 2026 | Alibaba | Qwen3.6-35B-A3B (open) | open-weights | – |
| 8 Apr 2026 | Meta | Muse Spark (proprietary; MSL) | flagship | [about.fb.com ✓ ↗](https://about.fb.com/news/2026/04/introducing-muse-spark-meta-superintelligence-labs/) |
| 7 Apr 2026 | Zhipu | GLM-5.1 | open-weights | [huggingface.co ↗](https://huggingface.co/zai-org/GLM-5.1) |
| 7 Apr 2026 | Anthropic | Claude Mythos Preview (restricted; Project Glasswing) | flagship | [anthropic.com ✓ ↗](https://www.anthropic.com/research/mythos-preview) |
| 2 Apr 2026 | Google | Gemma 4 | open-weights | [blog.google ✓ ↗](https://blog.google/innovation-and-ai/technology/developers-tools/gemma-4/) |
| 2 Apr 2026 | Alibaba | Qwen3.6-Plus | flagship | – |
| 18 Mar 2026 | MiniMax | MiniMax-M2.7 | open-weights | [minimax.io ✓ ↗](https://www.minimax.io/news/minimax-m27-en) |
| 16 Mar 2026 | Mistral | Mistral Small 4 | open-weights | [mistral.ai ✓ ↗](https://mistral.ai/news/mistral-small-4/) |
| 11 Mar 2026 | NVIDIA | Nemotron 3 Super | open-weights | [blogs.nvidia.com ✓ ↗](https://blogs.nvidia.com/blog/nemotron-3-super-agentic-ai/) |
| 5 Mar 2026 | OpenAI | GPT-5.4 (Thinking/Pro; mini/nano 03-17) | flagship | [openai.com ✓ ↗](https://openai.com/index/gpt-5-4-thinking-system-card/) |
| 3 Mar 2026 | Google | Gemini 3.1 Flash-Lite | small | [blog.google ✓ ↗](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-1-flash-lite/) |
| 19 Feb 2026 | Google | Gemini 3.1 Pro | flagship | [blog.google ✓ ↗](https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-1-pro/) |
| 17 Feb 2026 | xAI | Grok 4.20 (beta; GA/API 03-10) | flagship | [docs.x.ai ↗](https://docs.x.ai/developers/models/grok-4.20) |
| 17 Feb 2026 | Anthropic | Claude Sonnet 4.6 | flagship | [anthropic.com ✓ ↗](https://www.anthropic.com/news/claude-sonnet-4-6) |
| 16 Feb 2026 | Alibaba | Qwen3.5 (397B open + 3.5-Plus) | open-weights | [qwenlm.github.io ↗](https://qwenlm.github.io/qwen-code-docs/en/blog/updates/weekly-update-2026-03-03/) |
| 14 Feb 2026 | ByteDance | Seed 2.0 | flagship | [seed.bytedance.com ✓ ↗](https://seed.bytedance.com/en/blog/seed2-0-%E6%AD%A3%E5%BC%8F%E5%8F%91%E5%B8%83) |
| 12 Feb 2026 | OpenAI | GPT-5.3-Codex-Spark | small | [openai.com ✓ ↗](https://openai.com/index/introducing-gpt-5-3-codex-spark/) |
| 12 Feb 2026 | MiniMax | MiniMax-M2.5 | open-weights | [huggingface.co ✓ ↗](https://huggingface.co/MiniMaxAI/MiniMax-M2.5) |
| 11 Feb 2026 | Zhipu | GLM-5 | open-weights | [huggingface.co ↗](https://huggingface.co/zai-org/GLM-5) |
| 5 Feb 2026 | OpenAI | GPT-5.3-Codex | reasoning | [openai.com ↗](https://openai.com/ja-JP/index/introducing-gpt-5-3-codex/) |
| 5 Feb 2026 | Anthropic | Claude Opus 4.6 | flagship | [anthropic.com ✓ ↗](https://www.anthropic.com/news/claude-opus-4-6) |
| 2 Feb 2026 | StepFun | Step 3.5 Flash | open-weights | [huggingface.co ✓ ↗](https://huggingface.co/stepfun-ai/Step-3.5-Flash) |
| 27 Jan 2026 | Moonshot | Kimi K2.5 | open-weights | [kimi.com ↗](https://www.kimi.com/blog/kimi-k2-5?ref=upstract.com) |
| 26 Jan 2026 | Alibaba | Qwen3-Max-Thinking | reasoning | [qwen.ai ↗](https://qwen.ai/blog?id=qwen3-max-thinking&ref=testingcatalog.com) |
| 23 Dec 2025 | MiniMax | MiniMax-M2.1 | open-weights | [huggingface.co ↗](https://huggingface.co/MiniMaxAI/MiniMax-M2.1) |
| 22 Dec 2025 | Zhipu | GLM-4.7 | open-weights | [huggingface.co ✓ ↗](https://huggingface.co/zai-org/GLM-4.7) |
| 18 Dec 2025 | OpenAI | GPT-5.2-Codex | reasoning | [openai.com ✓ ↗](https://openai.com/index/introducing-gpt-5-2-codex/) |
| 17 Dec 2025 | Google | Gemini 3 Flash | small | [blog.google ✓ ↗](https://blog.google/products-and-platforms/products/gemini/gemini-3-flash/) |
| 16 Dec 2025 | Xiaomi | MiMo-V2-Flash | open-weights | [mimo.xiaomi.com ✓ ↗](https://mimo.xiaomi.com/mimo-v2-flash) |
| 15 Dec 2025 | NVIDIA | Nemotron 3 Nano | open-weights | [nvidianews.nvidia.com ✓ ↗](https://nvidianews.nvidia.com/news/nvidia-debuts-nemotron-3-family-of-open-models) |
| 11 Dec 2025 | OpenAI | GPT-5.2 | flagship | [openai.com ↗](https://openai.com/zh-Hans-CN/index/introducing-gpt-5-2/) |
| 9 Dec 2025 | Mistral | Devstral 2 (+ Devstral Small 2) | open-weights | [mistral.ai ✓ ↗](https://mistral.ai/news/devstral-2-vibe-cli/) |
| 4 Dec 2025 | Google | Gemini 3 Deep Think | reasoning | [blog.google ✓ ↗](https://blog.google/products-and-platforms/products/gemini/gemini-3-deep-think/) |
| 2 Dec 2025 | Mistral | Mistral Large 3 (+ Ministral 3) | open-weights | [mistral.ai ✓ ↗](https://mistral.ai/news/mistral-3/) |
| 2 Dec 2025 | Amazon | Nova 2 | flagship | [aboutamazon.com ✓ ↗](https://www.aboutamazon.com/news/aws/aws-re-invent-2025-ai-news-updates) |
| 1 Dec 2025 | DeepSeek | DeepSeek-V3.2 (+ Speciale) | open-weights | [api-docs.deepseek.com ↗](https://api-docs.deepseek.com/news/news251201/) |
| 27 Nov 2025 | DeepSeek | DeepSeekMath-V2 | open-weights | – |
| 24 Nov 2025 | Anthropic | Claude Opus 4.5 | flagship | [anthropic.com ✓ ↗](https://www.anthropic.com/news/claude-opus-4-5) |
| 19 Nov 2025 | xAI | Grok 4.1 Fast (+ Agent Tools API) | small | [x.ai ✓ ↗](https://x.ai/news/grok-4-1-fast) |
| 19 Nov 2025 | OpenAI | GPT-5.1-Codex-Max | reasoning | [openai.com ✓ ↗](https://openai.com/index/gpt-5-1-codex-max/) |
| 18 Nov 2025 | Google | Gemini 3 Pro | flagship | [blog.google ✓ ↗](https://blog.google/products-and-platforms/products/gemini/gemini-3/) |
| 17 Nov 2025 | xAI | Grok 4.1 | flagship | [x.ai ✓ ↗](https://x.ai/news/grok-4-1) |
| 13 Nov 2025 | Baidu | ERNIE 5.0 | flagship | – |
| 12 Nov 2025 | OpenAI | GPT-5.1 | flagship | [openai.com ✓ ↗](https://openai.com/index/gpt-5-system-card-addendum-gpt-5-1/) |
| 6 Nov 2025 | Moonshot | Kimi K2 Thinking | reasoning | [kimi.com ↗](https://www.kimi.com/en/blog/kimi-k2-thinking) |
| 30 Oct 2025 | Moonshot | Kimi Linear 48B-A3B | open-weights | [huggingface.co ↗](https://huggingface.co/moonshotai/Kimi-Linear-48B-A3B-Instruct) |
| 27 Oct 2025 | MiniMax | MiniMax-M2 | open-weights | – |
| 15 Oct 2025 | Anthropic | Claude Haiku 4.5 | small | [anthropic.com ✓ ↗](https://www.anthropic.com/news/claude-haiku-4-5) |
| 14 Oct 2025 | Ant Group | Ring-1T | open-weights | [huggingface.co ↗](https://huggingface.co/inclusionAI/Ring-1T) |
| 9 Oct 2025 | Ant Group | Ling-1T | open-weights | [huggingface.co ↗](https://huggingface.co/inclusionAI/Ling-1T) |
| 30 Sep 2025 | Zhipu | GLM-4.6 | open-weights | [huggingface.co ✓ ↗](https://huggingface.co/zai-org/GLM-4.6-FP8) |
| 29 Sep 2025 | DeepSeek | DeepSeek-V3.2-Exp (sparse attention) | open-weights | [api-docs.deepseek.com ↗](https://api-docs.deepseek.com/news/news250929/) |
| 29 Sep 2025 | Anthropic | Claude Sonnet 4.5 | flagship | [anthropic.com ✓ ↗](https://www.anthropic.com/news/claude-sonnet-4-5) |
| 19 Sep 2025 | xAI | Grok 4 Fast | small | [x.ai ✓ ↗](https://x.ai/news/grok-4-fast) |
| 15 Sep 2025 | OpenAI | GPT-5-Codex | reasoning | [openai.com ✓ ↗](https://openai.com/index/introducing-upgrades-to-codex/) |
| 11 Sep 2025 | Alibaba | Qwen3-Next-80B-A3B | open-weights | – |
| 5 Sep 2025 | Alibaba | Qwen3-Max (preview; official 09-24) | flagship | – |
| 1 Sep 2025 | Meituan | LongCat-Flash | open-weights | [huggingface.co ↗](https://huggingface.co/meituan-longcat/LongCat-Flash-Chat) |
| 28 Aug 2025 | xAI | Grok Code Fast 1 | small | [x.ai ✓ ↗](https://x.ai/news/grok-code-fast-1) |
| 28 Aug 2025 | Microsoft | MAI-1-preview | flagship | – |
| 23 Aug 2025 | xAI | Grok 2.5 open weights | open-weights | [techcrunch.com ↗](https://techcrunch.com/2025/08/24/elon-musk-says-xai-has-open-sourced-grok-2-5/) |
| 21 Aug 2025 | DeepSeek | DeepSeek-V3.1 (hybrid reasoning) | open-weights | [api-docs.deepseek.com ↗](https://api-docs.deepseek.com/news/news250821/) |
| 7 Aug 2025 | OpenAI | GPT-5 | flagship | [openai.com ✓ ↗](https://openai.com/index/introducing-gpt-5/) |
| 5 Aug 2025 | OpenAI | gpt-oss-120b/20b | open-weights | [openai.com ✓ ↗](https://openai.com/index/gpt-oss-model-card/) |
| 5 Aug 2025 | Anthropic | Claude Opus 4.1 | flagship | [anthropic.com ✓ ↗](https://www.anthropic.com/news/claude-opus-4-1) |
| 1 Aug 2025 | Google | Gemini 2.5 Deep Think | reasoning | [blog.google ✓ ↗](https://blog.google/products-and-platforms/products/gemini/gemini-2-5-deep-think/) |
| 28 Jul 2025 | Zhipu | GLM-4.5 | open-weights | [z.ai ✓ ↗](https://z.ai/blog/glm-4.5) |
| 25 Jul 2025 | StepFun | Step 3 | open-weights | – |
| 22 Jul 2025 | Alibaba | Qwen3-Coder | open-weights | [qwenlm.github.io ✓ ↗](https://qwenlm.github.io/blog/qwen3-coder/) |
| 11 Jul 2025 | Moonshot | Kimi K2 | open-weights | [kimi.com ↗](https://www.kimi.com/en/blog/kimi-k2) |
| 9 Jul 2025 | xAI | Grok 4 (+ Heavy) | flagship | [x.ai ✓ ↗](https://x.ai/news/grok-4) |
| 27 Jun 2025 | Tencent | Hunyuan-A13B | open-weights | [huggingface.co ✓ ↗](https://huggingface.co/tencent/Hunyuan-A13B-Instruct) |
| 17 Jun 2025 | Google | Gemini 2.5 Flash-Lite (preview; 2.5 Pro/Flash GA) | small | [developers.googleblog.com ✓ ↗](https://developers.googleblog.com/en/gemini-2-5-thinking-model-updates/) |
| 16 Jun 2025 | MiniMax | MiniMax-M1 | open-weights | [huggingface.co ✓ ↗](https://huggingface.co/collections/MiniMaxAI/minimax-m1-68502ad9634ec0eeac8cf094) |
| 11 Jun 2025 | ByteDance | Seed 1.6 | reasoning | – |
| 10 Jun 2025 | OpenAI | o3-pro | reasoning | [openai.com ✓ ↗](https://openai.com/index/introducing-o3-and-o4-mini/) |
| 10 Jun 2025 | Mistral | Magistral (Small/Medium) | reasoning | [mistral.ai ✓ ↗](https://mistral.ai/news/magistral/) |
| 28 May 2025 | DeepSeek | DeepSeek-R1-0528 | reasoning | [api-docs.deepseek.com ↗](https://api-docs.deepseek.com/news/news250528/) |
| 22 May 2025 | Anthropic | Claude Opus 4 + Sonnet 4 | flagship | [anthropic.com ✓ ↗](https://www.anthropic.com/news/claude-4) |
| 21 May 2025 | Mistral | Devstral | open-weights | [mistral.ai ✓ ↗](https://mistral.ai/news/devstral/) |
| 20 May 2025 | Google | Gemma 3n (preview; full 06-26) | small | – |
| 16 May 2025 | OpenAI | codex-1 (Codex research preview) | reasoning | [openai.com ✓ ↗](https://openai.com/index/introducing-codex/) |
| 7 May 2025 | Mistral | Mistral Medium 3 | flagship | [mistral.ai ✓ ↗](https://mistral.ai/news/mistral-medium-3/) |
| 30 Apr 2025 | Xiaomi | MiMo-7B | open-weights | [huggingface.co ✓ ↗](https://huggingface.co/XiaomiMiMo/MiMo-7B-RL) |
| 30 Apr 2025 | Microsoft | Phi-4-reasoning | open-weights | [huggingface.co ✓ ↗](https://huggingface.co/microsoft/Phi-4-reasoning) |
| 30 Apr 2025 | Amazon | Nova Premier | flagship | [aws.amazon.com ✓ ↗](https://aws.amazon.com/ru/blogs/aws/amazon-nova-premier-our-most-capable-model-for-complex-tasks-and-teacher-for-model-distillation/) |
| 29 Apr 2025 | Alibaba | Qwen3 | open-weights | [qwenlm.github.io ✓ ↗](https://qwenlm.github.io/blog/qwen3/) |
| 17 Apr 2025 | Google | Gemini 2.5 Flash | small | [blog.google ✓ ↗](https://blog.google/products-and-platforms/products/gemini/gemini-2-5-flash-preview/) |
| 16 Apr 2025 | OpenAI | o3 + o4-mini | reasoning | [openai.com ✓ ↗](https://openai.com/index/introducing-o3-and-o4-mini/) |
| 14 Apr 2025 | OpenAI | GPT-4.1 | flagship | [en.wikipedia.org ↗](https://en.wikipedia.org/wiki/GPT-4.1) |
| 5 Apr 2025 | Meta | Llama 4 (Scout/Maverick) | open-weights | [about.fb.com ✓ ↗](https://about.fb.com/ltam/news/2025/04/la-coleccion-de-modelos-llama-4-el-inicio-de-una-nueva-era-de-innovacion-multimodal-nativa-para-inteligencia-artificial/) |
| 25 Mar 2025 | Google | Gemini 2.5 Pro | reasoning | [blog.google ✓ ↗](https://blog.google/innovation-and-ai/models-and-research/google-deepmind/gemini-model-thinking-updates-march-2025/) |
| 21 Mar 2025 | Tencent | Hunyuan-T1 | reasoning | – |
| 16 Mar 2025 | Baidu | ERNIE X1 | reasoning | – |
| 16 Mar 2025 | Baidu | ERNIE 4.5 | flagship | – |
| 12 Mar 2025 | Google | Gemma 3 | open-weights | [blog.google ✓ ↗](https://blog.google/innovation-and-ai/technology/developers-tools/gemma-3/) |
| 6 Mar 2025 | Alibaba | QwQ-32B | reasoning | [qwenlm.github.io ✓ ↗](https://qwenlm.github.io/blog/qwq-32b/) |
| 27 Feb 2025 | OpenAI | GPT-4.5 | flagship | [en.wikipedia.org ↗](https://en.wikipedia.org/wiki/GPT-4.5) |
| 24 Feb 2025 | Anthropic | Claude 3.7 Sonnet | reasoning | [anthropic.com ✓ ↗](https://www.anthropic.com/news/claude-3-7-sonnet) |
| 17 Feb 2025 | xAI | Grok 3 (+ Grok 3 mini) | flagship | [x.ai ✓ ↗](https://x.ai/news/grok-3) |
| 5 Feb 2025 | Google | Gemini 2.0 Pro (exp) + 2.0 Flash GA | flagship | [blog.google ✓ ↗](https://blog.google/innovation-and-ai/models-and-research/google-deepmind/gemini-model-updates-february-2025/) |
| 31 Jan 2025 | OpenAI | o3-mini | reasoning | [openai.com ✓ ↗](https://openai.com/index/openai-o3-mini/) |
| 30 Jan 2025 | Mistral | Mistral Small 3 | open-weights | [mistral.ai ✓ ↗](https://mistral.ai/news/mistral-small-3/) |
| 28 Jan 2025 | Alibaba | Qwen2.5-Max | flagship | [qwenlm.github.io ✓ ↗](https://qwenlm.github.io/blog/qwen2.5-max/) |
| 26 Jan 2025 | Alibaba | Qwen2.5-VL | open-weights | [qwenlm.github.io ✓ ↗](https://qwenlm.github.io/blog/qwen2.5-vl/) |
| 22 Jan 2025 | ByteDance | Doubao 1.5 Pro | flagship | – |
| 20 Jan 2025 | Moonshot | Kimi k1.5 | reasoning | [kimi.com ✓ ↗](https://www.kimi.com/en/blog/) |
| 20 Jan 2025 | DeepSeek | DeepSeek-R1 | reasoning | [api-docs.deepseek.com ↗](https://api-docs.deepseek.com/news/news250120/) |
| 15 Jan 2025 | MiniMax | MiniMax-01 | open-weights | [huggingface.co ↗](https://huggingface.co/MiniMaxAI/MiniMax-Text-01) |
| 26 Dec 2024 | DeepSeek | DeepSeek-V3 | open-weights | [api-docs.deepseek.com ↗](https://api-docs.deepseek.com/news/news1226/) |
| 20 Dec 2024 | OpenAI | o3 / o3-mini (announced; preview only) | reasoning | [en.wikipedia.org ↗](https://en.wikipedia.org/wiki/OpenAI_o3) |
| 19 Dec 2024 | Google | Gemini 2.0 Flash Thinking (exp) | reasoning | [ai.google.dev ✓ ↗](https://ai.google.dev/gemini-api/docs/changelog) |
| 12 Dec 2024 | Microsoft | Phi-4 | open-weights | – |
| 11 Dec 2024 | Google | Gemini 2.0 Flash (experimental) | flagship | [blog.google ✓ ↗](https://blog.google/innovation-and-ai/models-and-research/google-deepmind/google-gemini-ai-update-december-2024/) |
| 6 Dec 2024 | Meta | Llama 3.3 70B | open-weights | [about.fb.com ↗](https://about.fb.com/ltam/news/2024/12/el-futuro-de-la-ia-construido-con-llama/) |
| 5 Dec 2024 | OpenAI | o1 (full) + o1 pro | reasoning | – |
| 3 Dec 2024 | Amazon | Amazon Nova | flagship | – |
| 28 Nov 2024 | Alibaba | QwQ-32B-Preview | reasoning | [qwen.ai ↗](https://qwen.ai/blog?id=qwq-32b-preview) |
| 20 Nov 2024 | DeepSeek | DeepSeek-R1-Lite-Preview | reasoning | [deepseek.com ✓ ↗](https://www.deepseek.com/en/news/r1-lite-preview/) |
| 12 Nov 2024 | Alibaba | Qwen2.5-Coder-32B | open-weights | [qwenlm.github.io ✓ ↗](https://qwenlm.github.io/blog/qwen2.5-coder-family/) |
| 5 Nov 2024 | Tencent | Hunyuan-Large | open-weights | – |
| 22 Oct 2024 | Anthropic | Claude 3.5 Sonnet (new) + 3.5 Haiku (avail. 11-04) | flagship | [anthropic.com ✓ ↗](https://www.anthropic.com/news/3-5-models-and-computer-use) |
| 16 Oct 2024 | Mistral | Ministral 3B/8B | small | [mistral.ai ✓ ↗](https://mistral.ai/news/ministraux/) |
| 15 Oct 2024 | NVIDIA | Llama-3.1-Nemotron-70B | open-weights | – |
| 25 Sep 2024 | Meta | Llama 3.2 (1B/3B text + 11B/90B vision) | open-weights | [ai.meta.com ✓ ↗](https://ai.meta.com/blog/llama-3-2-connect-2024-Vision-edge-mobile-devices/) |
| 19 Sep 2024 | Alibaba | Qwen2.5 | open-weights | [qwenlm.github.io ✓ ↗](https://qwenlm.github.io/blog/qwen2.5/) |
| 12 Sep 2024 | OpenAI | o1-preview / o1-mini | reasoning | [openai.com ✓ ↗](https://openai.com/index/introducing-openai-o1-preview/) |
| 11 Sep 2024 | Mistral | Pixtral 12B | open-weights | [mistral.ai ↗](https://mistral.ai/news/pixtral-12b/) |
| 13 Aug 2024 | xAI | Grok-2 | flagship | [x.ai ✓ ↗](https://x.ai/news/grok-2) |
| 24 Jul 2024 | Mistral | Mistral Large 2 | flagship | [mistral.ai ✓ ↗](https://mistral.ai/news/mistral-large-2407/) |
| 23 Jul 2024 | Meta | Llama 3.1 (405B) | open-weights | [ai.meta.com ✓ ↗](https://ai.meta.com/blog/meta-llama-3-1/) |
| 18 Jul 2024 | OpenAI | GPT-4o mini | small | [developers.openai.com ↗](https://developers.openai.com/api/docs/models/gpt-4o-mini-tts) |
| 27 Jun 2024 | Google | Gemma 2 | open-weights | [blog.google ✓ ↗](https://blog.google/innovation-and-ai/technology/developers-tools/google-gemma-2/) |
| 20 Jun 2024 | Anthropic | Claude 3.5 Sonnet | flagship | [anthropic.com ✓ ↗](https://www.anthropic.com/news/claude-3-5-sonnet) |
| 17 Jun 2024 | DeepSeek | DeepSeek-Coder-V2 | open-weights | [api-docs.deepseek.com ↗](https://api-docs.deepseek.com/updates/) |
| 14 Jun 2024 | NVIDIA | Nemotron-4 340B | open-weights | [blogs.nvidia.com ✓ ↗](https://blogs.nvidia.com/blog/nemotron-4-synthetic-data-generation-llm-training/) |
| 7 Jun 2024 | Alibaba | Qwen2 | open-weights | [qwenlm.github.io ✓ ↗](https://qwenlm.github.io/blog/qwen2/) |
| 29 May 2024 | Mistral | Codestral | open-weights | [mistral.ai ✓ ↗](https://mistral.ai/news/codestral/) |
| 15 May 2024 | ByteDance | Doubao Pro | flagship | – |
| 14 May 2024 | Google | Gemini 1.5 Flash | small | [blog.google ✓ ↗](https://blog.google/technology/ai/google-gemini-update-flash-ai-assistant-io-2024/) |
| 13 May 2024 | OpenAI | GPT-4o | flagship | [openai.com ✓ ↗](https://openai.com/index/hello-gpt-4o/) |
| 6 May 2024 | DeepSeek | DeepSeek-V2 | open-weights | [github.com ✓ ↗](https://github.com/deepseek-ai/DeepSeek-V2) |
| 23 Apr 2024 | Microsoft | Phi-3 | open-weights | [huggingface.co ✓ ↗](https://huggingface.co/microsoft/Phi-3-mini-4k-instruct) |
| 18 Apr 2024 | Meta | Llama 3 | open-weights | [ai.meta.com ✓ ↗](https://ai.meta.com/blog/meta-llama-3/) |
| 10 Apr 2024 | Mistral | Mixtral 8x22B | open-weights | [mistral.ai ↗](https://mistral.ai/news/mixtral-8x22b/) |
| 28 Mar 2024 | xAI | Grok-1.5 | flagship | [x.ai ✓ ↗](https://x.ai/news/grok-1.5) |
| 17 Mar 2024 | xAI | Grok-1 open weights | open-weights | [x.ai ✓ ↗](https://x.ai/news/grok-os) |
| 4 Mar 2024 | Anthropic | Claude 3 (Opus/Sonnet; Haiku 03-13) | flagship | [anthropic.com ✓ ↗](https://www.anthropic.com/news/claude-3-family) |
| 26 Feb 2024 | Mistral | Mistral Large | flagship | [mistral.ai ✓ ↗](https://mistral.ai/news/mistral-large/) |
| 21 Feb 2024 | Google | Gemma (2B/7B) | open-weights | [blog.google ✓ ↗](https://blog.google/innovation-and-ai/technology/developers-tools/gemma-open-models/) |
| 15 Feb 2024 | Google | Gemini 1.5 Pro | flagship | [blog.google ✓ ↗](https://blog.google/innovation-and-ai/products/google-gemini-next-generation-model-february-2024/) |
| 8 Feb 2024 | Google | Gemini 1.0 Ultra (Gemini Advanced) | flagship | [blog.google ↗](https://blog.google/innovation-and-ai/products/google-gemini-next-generation-model-february-2024/) |
| 4 Feb 2024 | Alibaba | Qwen1.5 | open-weights | [qwenlm.github.io ✓ ↗](https://qwenlm.github.io/blog/qwen1.5/) |
| 16 Jan 2024 | Zhipu | GLM-4 | flagship | – |
| 12 Dec 2023 | Microsoft | Phi-2 | open-weights | [huggingface.co ✓ ↗](https://huggingface.co/microsoft/phi-2) |
| 8 Dec 2023 | Mistral | Mixtral 8x7B | open-weights | [mistral.ai ↗](https://mistral.ai/news/mixtral-of-experts/) |
| 6 Dec 2023 | Google | Gemini 1.0 (Pro/Nano; Ultra announced) | flagship | [blog.google ✓ ↗](https://blog.google/innovation-and-ai/technology/ai/google-gemini-ai/) |
| 30 Nov 2023 | Alibaba | Qwen-72B | open-weights | [github.com ✓ ↗](https://github.com/QwenLM/Qwen) |
| 29 Nov 2023 | DeepSeek | DeepSeek LLM 67B | open-weights | – |
| 21 Nov 2023 | Anthropic | Claude 2.1 | flagship | [anthropic.com ✓ ↗](https://www.anthropic.com/news/claude-2-1) |
| 6 Nov 2023 | OpenAI | GPT-4 Turbo | flagship | [developers.openai.com ↗](https://developers.openai.com/api/docs/models/gpt-4-turbo) |
| 4 Nov 2023 | xAI | Grok-1 | flagship | [en.wikipedia.org ↗](https://en.wikipedia.org/wiki/Grok_(chatbot)) |
| 2 Nov 2023 | DeepSeek | DeepSeek Coder | open-weights | – |
| 17 Oct 2023 | Baidu | ERNIE 4.0 | flagship | – |
| 9 Oct 2023 | Moonshot | Kimi Chat (debut; 200K-char context) | flagship | [kimi.com ↗](https://www.kimi.com/en/help/others/chat-issues) |
| 27 Sep 2023 | Mistral | Mistral 7B | open-weights | [mistral.ai ✓ ↗](https://mistral.ai/news/announcing-mistral-7b/) |
| 24 Aug 2023 | Meta | Code Llama | open-weights | [ai.meta.com ✓ ↗](https://ai.meta.com/blog/code-llama-large-language-model-coding/) |
| 9 Aug 2023 | Anthropic | Claude Instant 1.2 | small | [anthropic.com ✓ ↗](https://www.anthropic.com/news/releasing-claude-instant-1-2) |
| 3 Aug 2023 | Alibaba | Qwen-7B (first open weights) | open-weights | – |
| 18 Jul 2023 | Meta | Llama 2 | open-weights | [ai.meta.com ✓ ↗](https://ai.meta.com/research/publications/llama-2-open-foundation-and-fine-tuned-chat-models/) |
| 11 Jul 2023 | Anthropic | Claude 2 | flagship | [anthropic.com ✓ ↗](https://www.anthropic.com/news/claude-2) |
| 10 May 2023 | Google | PaLM 2 | flagship | [blog.google ✓ ↗](https://blog.google/innovation-and-ai/products/google-palm-2-ai-large-language-model/) |
| 11 Apr 2023 | Alibaba | Qwen / Tongyi Qianwen (beta) | flagship | [en.wikipedia.org ↗](https://en.wikipedia.org/wiki/Qwen) |
| 14 Mar 2023 | Zhipu | ChatGLM-6B | open-weights | [huggingface.co ✓ ↗](https://huggingface.co/zai-org/chatglm-6b) |
| 14 Mar 2023 | OpenAI | GPT-4 | flagship | [openai.com ✓ ↗](https://openai.com/index/gpt-4-research/) |
| 14 Mar 2023 | Anthropic | Claude (1) | flagship | [anthropic.com ✓ ↗](https://www.anthropic.com/news/introducing-claude) |
| 24 Feb 2023 | Meta | LLaMA (1) | open-weights | [en.wikipedia.org ↗](https://en.wikipedia.org/wiki/Llama_(language_model)) |
| 30 Nov 2022 | OpenAI | ChatGPT (GPT-3.5) | flagship | [openai.com ✓ ↗](https://openai.com/index/chatgpt/) |

---
Source: https://themodelindex.org/data/ · The Model Index · data as of 2026-10-01