| Qwen3 | Qwen | 61 | 27 Apr 2025 · 14 May 2026 | 595.8M – 480B | embeddingimagescoretext | Apache-2.0 10MIT 1 | Aider polyglot #20SWE-bench Verified #25Aider polyglot — well-formed responses #45 |
| DeepSeek | DeepSeek | 50 | 29 Oct 2023 · 10 Sept 2026 | 3.34B – 1.6T | imagetext | MIT 11Other 4Apache-2.0 1 | LiveBench Agentic Coding #1LiveBench #5Aider polyglot #8 |
| Qwen2.5 | Qwen Team | 47 | 16 Sept 2024 · 27 Mar 2025 | 494M – 72.7B | audioimagetextvideo | Apache-2.0 5Research-Only 2Other 1 | Aider polyglot — well-formed responses #1SWE-bench Verified #42Aider polyglot #45 |
| Mistral | Mistral AI | 46 | 20 Sept 2023 · 23 Jun 2026 | 7B – 127.7B | documentimagetext | Apache-2.0 12MIT-Modified 1CC-BY-NC-4.0 1+1 | τ²-bench #31Terminal-Bench #67IFBench #68 |
| GPT 4 | OpenAI | 30 | 28 May 2023 · 20 Nov 2024 | — | imagetext | — | Aider polyglot — well-formed responses #1SWE-bench (full test split) #3SWE-bench Lite #3 |
| Qwen | Qwen Team | 26 | 15 Nov 2024 · 8 Sept 2025 | 1.78B – 72B | audioimagetextvideo | MIT 3Apache-2.0 2 | Aider polyglot #48Aider polyglot — well-formed responses #54IFBench #210 |
| Command | Cohere | 22 | 11 Mar 2024 · 28 Jul 2025 | 8.03B – 111.9B | imagetext | CC-BY-NC-4.0 5 | Aider polyglot — well-formed responses #11IFBench #34Aider polyglot #56 |
| MiniMax | MiniMax | 22 | 12 Jan 2025 · 7 Aug 2026 | 857.9M – 456.4B | imagetextvideo | Other 13Apache-2.0 5 | SWE-bench Verified #2IFBench #3SWE-bench Multilingual #6 |
| Claude | Anthropic | 20 | 22 May 2025 · 1 Sept 2026 | — | imagetext | — | AA Intelligence Index #1Humanity's Last Exam #1LiveBench #1 |
| Gemini 2.5 | Google | 20 | 7 May 2025 · 7 Oct 2025 | — | audiodocumentimagetextvideo | — | Aider polyglot — well-formed responses #1Aider polyglot #3SWE-bench Verified #29 |
| Llama 3.1 | Meta AI | 18 | 14 Jul 2024 · 28 Aug 2024 | 8.03B – 405.9B | text | Llama-3.1-Community 7Llama-3-Community 1 | Terminal-Bench #167GPQA Diamond #205IFBench #208 |
| Kimi | Moonshot AI | 17 | 9 Apr 2025 · 13 Jun 2026 | 9.77B – 2.78T | imagetextvideo | MIT 9Other 8 | LiveBench Reasoning #5SciCode #5SWE-bench Multilingual #7 |
| LFM2.5 | Liquid AI | 17 | 5 Jan 2026 · 11 Aug 2026 | 229.7M – 8.47B | — | Other 8 | IFBench #120SciCode #126MMMU-Pro #152 |
| Embed | Cohere | 15 | 2 Nov 2023 · 13 Mar 2026 | 595.8M – 4.02B | embeddingimagetext | MIT 5 | no results |
| Qwen3.5 | Qwen | 15 | 16 Feb 2026 · 27 Apr 2026 | 4.66B – 9.65B | imagetextvideo | Apache-2.0 2 | IFBench #11τ²-bench #18MMMU-Pro #32 |
| ERNIE 4.5 | Baidu | 14 | 28 Jun 2025 · 7 Nov 2025 | 360.8M – 423.5B | imagetext | Apache-2.0 13 | GPQA Diamond #134Terminal-Bench #196IFBench #207 |
| FLUX | Black Forest Labs | 14 | 31 Jul 2024 · 6 Apr 2026 | 62.4M – 32.2B | — | Other 10Apache-2.0 4 | no results |
| Granite | IBM | 13 | 8 Oct 2024 · 20 Apr 2026 | 805.3K – 8.17B | — | Apache-2.0 13 | no results |
| Llama 3 | Nous Research | 13 | 17 Apr 2024 · 13 Aug 2024 | 8.03B – 70.5B | text | Llama-3-Community 4Other 3Apache-2.0 1 | IFBench #222Terminal-Bench #262τ²-bench #302 |
| Whisper | OpenAI | 13 | 26 Sept 2022 · 1 Oct 2024 | 37.8M – 1.54B | — | Apache-2.0 11MIT 1 | no results |
| Aya | Cohere | 12 | 23 Oct 2024 · 13 Feb 2026 | 3.35B – 32.3B | imagetext | CC-BY-NC-4.0 5 | Terminal-Bench #276Humanity's Last Exam #296τ²-bench #302 |
| Gemma 3 | Google | 12 | 20 Feb 2025 · 13 Mar 2025 | 999.9M – 4.3B | imagetext | Gemma-Terms 2 | Aider polyglot — well-formed responses #1Aider polyglot #60SciCode #120 |
| Grok | xAI | 12 | 20 May 2026 · 20 May 2026 | — | audioimagetextvideo | — | MMMU-Pro #35LiveBench Instruction Following #37Humanity's Last Exam #40 |
| Jamba | AI21 Labs | 12 | 28 Mar 2024 · 5 Oct 2025 | 127.7M – 398.6B | — | Other 6Apache-2.0 5 | IFBench #135Terminal-Bench #262τ²-bench #268 |
| Llama | Meta AI | 12 | 9 Aug 2024 · 17 Jul 2026 | 68M – 8.03B | imagetext | CC-BY-NC-4.0 1Other 1 | SWE-bench (full test split) #4SWE-bench Lite #4GPQA Diamond #189 |
| OpenAI o-series | OpenAI | 12 | 17 Dec 2024 · 10 Jun 2025 | — | imagetext | — | SWE-bench Multimodal #1Aider polyglot #2SWE-bench Verified #22 |
| GPT 5 | OpenAI | 11 | 7 Aug 2025 · 16 Oct 2025 | — | imagetext | — | Aider polyglot #1SWE-bench Multilingual #13SWE-bench Verified #17 |
| Llama 2 | Meta AI | 11 | 13 Jul 2023 · 8 Aug 2023 | 6.74B – 13B | — | Llama-2-Community 5MIT 1 | Humanity's Last Exam #291AA Intelligence Index #402GPQA Diamond #406 |
| CLIP | Apple | 10 | 2 Mar 2022 · 30 Mar 2026 | 427.6M – 39B | audioimagetext | Apple-AMLR 5 | no results |
| GPT2 | Stanford CRFM | 10 | 2 Mar 2022 · 2 Mar 2022 | — | — | Apache-2.0 1 | no results |
| Gemini | Google | 10 | — · — | — | audiodocumentimagetextvideo | — | Aider polyglot — well-formed responses #20Aider polyglot #38 |
| Gemini 2.0 | Google | 10 | — · — | — | audioimagetextvideo | — | Aider polyglot — well-formed responses #1Aider polyglot #39SWE-bench Verified #40 |
| Llama 3.2 | Meta AI | 10 | 18 Sept 2024 · 27 Sept 2024 | 1.24B – 88.6B | text | Llama-3.2-Community 7 | MMMU-Pro #142τ²-bench #245Terminal-Bench #262 |
| Ministral 3 | Mistral AI | 10 | 31 Oct 2025 · 2 Dec 2025 | 3.85B – 13.9B | imagetext | Apache-2.0 7 | SciCode #119MMMU-Pro #129τ²-bench #207 |
| Qwen2 | Qwen Team | 10 | 9 Aug 2024 · 29 Aug 2024 | 7B | audioimagetextvideo | Apache-2.0 1 | AA Intelligence Index #374GPQA Diamond #384Humanity's Last Exam #413 |
| Gemini 3.1 | Google | 9 | 19 Feb 2026 · 30 Jun 2026 | — | audiodocumentimagetextvideo | — | LiveBench Instruction Following #3GPQA Diamond #5MMMU-Pro #8 |
| InternLM2 | InternLM (Shanghai AI Laboratory) | 9 | 10 Jan 2024 · 31 Jul 2024 | 1.89B – 20B | — | Other 8Apache-2.0 1 | no results |
| Nova | Amazon Web Services | 9 | 5 Dec 2024 · 31 Oct 2025 | — | imagetext | — | MMMU-Pro #138τ²-bench #165Terminal-Bench #184 |
| SmolLM | Hugging Face | 9 | 14 Jul 2024 · 20 Jan 2025 | 134.5M – 2.25B | — | Apache-2.0 8 | no results |
| Solar | Upstage | 9 | 1 Jan 2024 · 10 Aug 2026 | 10.7B | text | Apache-2.0 1 | SciCode #46IFBench #55AA Intelligence Index #56 |
| T5 | Amazon Web Services | 9 | 2 Mar 2022 · 28 Feb 2024 | 8.39M – 4.02B | — | Apache-2.0 6MIT 2 | no results |
| Devstral | Mistral AI | 8 | 28 Nov 2025 · 28 Nov 2025 | 24B | — | Apache-2.0 1 | SWE-bench Verified #23SciCode #105MMMU-Pro #136 |
| Gemma 4 | Google | 8 | 2 Mar 2026 · 23 May 2026 | 5.12B – 31.3B | imagetextvideo | Apache-2.0 5 | IFBench #27Terminal-Bench #49SciCode #55 |
| Magistral | Mistral AI | 8 | — · — | — | — | — | MMMU-Pro #109τ²-bench #142Terminal-Bench #152 |
| Seed | ByteDance | 8 | 27 Apr 2025 · 20 Aug 2025 | 7.51B – 36.1B | — | MIT 3Apache-2.0 2Other 1 | τ²-bench #69MMMU-Pro #81Terminal-Bench #94 |
| Voxtral | Mistral AI | 8 | 1 Jul 2025 · 21 Jan 2026 | 4.43B – 24.3B | audiotext | Apache-2.0 5 | no results |
| GPT 5.6 | OpenAI | 7 | 9 Jul 2026 · 9 Jul 2026 | — | imagetext | — | Terminal-Bench #1LiveBench Coding #3LiveBench Mathematics #3 |
| Gemini 3 | Google | 7 | 20 Nov 2025 · 18 Jun 2026 | — | audiodocumentimagetextvideo | — | SWE-bench Multilingual #1SWE-bench Verified #2IFBench #12 |
| Granite 4.0 | IBM | 7 | 20 Oct 2025 · 20 Oct 2025 | — | text | — | τ²-bench #234Terminal-Bench #237Humanity's Last Exam #252 |
| Mixtral 8 | Mistral AI | 7 | 10 Dec 2023 · 17 Apr 2024 | 46.7B – 140.6B | text | Apache-2.0 3 | Humanity's Last Exam #332AA Intelligence Index #401GPQA Diamond #401 |