Share
LLM Agent Leaderboard
As of
2026-07-17
| Rank | Name | Description |
|---|---|---|
| 🥇 | Claude Fable 5 (High) | By Anthropic · net improvement 13.94% (+0.60%p MoM, holds #1) |
| 🥈 | GPT 5.6 Sol (xHigh) | By OpenAI · net improvement 10.94% (new model, added Jul 13 · debuts at #2) |
| 🥉 | Claude Opus 4.8 (Thinking) | By Anthropic · net improvement 9.28% (-0.09%p MoM, #2 to #3) |
| 4. | GPT 5.5 (xHigh) | By OpenAI · net improvement 8.26% (+0.05%p MoM, #3 to #4) |
| 5. | Claude Sonnet 5 (High) | By Anthropic · net improvement 8.00% (new model · debuts at #5) |
| 6. | Claude Opus 4.7 (Thinking) | By Anthropic · net improvement 7.73% (-0.34%p MoM, #5 to #6) |
| 7. | Claude Opus 4.7 | By Anthropic · net improvement 7.63% (-0.53%p MoM, #4 to #7) |
| 8. | GPT 5.5 (High) | By OpenAI · net improvement 7.16% (+0.03%p MoM, #6 to #8) |
| 9. | GLM 5.2 (Max) | By Z.ai · net improvement 6.24% (-0.69%p MoM, #7 to #9) |
| 10. | GPT 5.5 | By OpenAI · net improvement 6.05% (-0.17%p MoM, holds #10) |
| 11. | Claude Opus 4.6 | By Anthropic · net improvement 5.88% (-0.59%p MoM, #9 to #11) |
| 12. | GPT 5.4 (High) | By OpenAI · net improvement 5.86% (-0.79%p MoM, #8 to #12) |
| 13. | Grok 4.5 | By SpaceXAI · net improvement 4.99% (new model, added Jul 13) |
| 14. | Claude Opus 4.8 | By Anthropic · net improvement 3.90% (newly shown with the top-20 expansion) |
| 15. | Claude Sonnet 4.6 | By Anthropic · net improvement 1.99% (newly shown with the top-20 expansion) |
| 16. | GLM 5.1 | By Z.ai · net improvement 1.19% (newly shown with the top-20 expansion) |
| 17. | Muse Spark 1.1 | By Meta · net improvement 0.17% (new model, added Jul 13) |
| 18. | Qwen3.7 Plus | By Alibaba · net improvement 0.61% (newly shown with the top-20 expansion) |
| 19. | Gemini 3.1 Pro Preview | By Google · net improvement 0.70% (newly shown with the top-20 expansion) |
| 20. | Kimi K2.7 Code | By Moonshot · net improvement 0.76% (newly shown with the top-20 expansion) |
This page compiles publicly available facts with clear source attribution. If you believe it infringes your rights, please contact us and we'll review and act promptly.