LLM benchmarks
Every model's score on every benchmark we track, each linked to where it was published. Independent runs are listed apart from the numbers labs report for their own models. Many of them feed the Respan Index.
Every model's score on every benchmark we track, each linked to where it was published. Independent runs are listed apart from the numbers labs report for their own models. Many of them feed the Respan Index.
Every model's score on every benchmark we track, each linked to where it was published. Independent runs are listed apart from the numbers labs report for their own models. Many of them feed the Respan Index.
Every model's score on every benchmark we track, each linked to where it was published. Independent runs are listed apart from the numbers labs report for their own models. Many of them feed the Respan Index.
Models ranked on one 0 to 100 score from public benchmarks and lab results.
169 ranked modelsVersion 1.1How it is scored
| Model | Respan Index | |
|---|---|---|
| 1. | Claude Fable 5.2by Anthropic | 81.0 |
| 2. | Gemini 4 Carbonby Google | 76.3 |
| 3. | Claude Opus 5.5by Anthropic | 76.2 |
| 4. | Claude Sonnet 5.5by Anthropic | 73.4 |
| 5. | Claude Fable 5.1by Anthropic | 72.4 |
| 5. | GPT-6 Astraby OpenAI | 72.4 |
| 7. | Kimi K3.1 Proby Moonshot AI | 72.0 |
| 8. | Muse Wonder Fullby Meta | 71.5 |
| 9. | GPT-6.1 Solby OpenAI | 69.9 |
| 10. | Gemini 4 Argonby Google | 69.2 |
| 10. | GPT-6.1 Terraby OpenAI | 69.2 |
| 12. | Claude Opus 5by Anthropic | 67.1 |
| 13. | Claude Fable 5by Anthropic | 63.9 |
| 14. | GPT-6 Solby OpenAI | 63.0 |
| 15. | Muse Emberby Meta | 61.9 |
| 16. | GPT-5.6 Solby OpenAI | 61.8 |
| 17. | Muse Wonderby Meta | 60.9 |
| 18. | DeepSeek-V4.1-Flash-1017by DeepSeek | 59.5 |
| 19. | Claude Haiku 5.5by Anthropic | 56.8 |
| 20. | Step 5 Previewby StepFun | 56.0 |
| 21. | Muse Spark 1.3by Meta | 55.8 |
| 21. | MiMo-V2.6-Proby Xiaomi MiMo | 55.8 |
| 23. | Claude Opus 4.8by Anthropic | 55.7 |
| 24. | GPT-5.5by OpenAI | 55.3 |
| 25. | GPT-5.6 Terraby OpenAI | 55.2 |
* Estimated, no published scores in that category.
A human-preference leaderboard: people compare two anonymous models' answers blind and vote, producing an Elo rating from millions of pairwise votes.
Arena176 modelsRun by huggingface.co/lmarena-ai, Nov 9, 2026