Vai al contenuto
AI.info

Model Tracker

Vision Arena leaderboard

The models in the Model Tracker that Arena has measured on Vision Arena, ranked by the highest score of each: the name Arena gives it, the setting it ran and the day. Arena’s leaderboard dataset (lmarena-ai/leaderboard-dataset on Hugging Fa

In inglese

The models in the Model Tracker that Arena has measured on Vision Arena, ranked by the highest score of each: the name Arena gives it, the setting it ran and the day. Arena’s leaderboard dataset (lmarena-ai/leaderboard-dataset on Hugging Face), CC BY 4.0: ratings from people’s votes between two anonymous models, as Arena publishes them.

  1. 1. Claude Fable 5: 1,308 (1,301–1,315) — claude-fable-5-high, high, 9 October 2026
  2. 2. Qwen3.8-Max: 1,301 (1,293–1,308) — qwen3.8-max, 9 October 2026
  3. 3. Claude Opus 4.6: 1,299 (1,293–1,306) — claude-opus-4-6-high, high, 9 October 2026
  4. 4. Claude Opus 4.7: 1,298 (1,292–1,305) — claude-opus-4-7, 9 October 2026
  5. 5. Gemini 3.7 Flash: 1,297 (1,287–1,307) — gemini-3.7-flash-high, high, 9 October 2026
  6. 6. GPT-6.1 Sol: 1,294 (1,280–1,308) — gpt-6.1-sol-max, max, 9 October 2026
  7. 7. Muse Spark: 1,294 (1,285–1,303) — muse-spark, 9 October 2026
  8. 8. Claude Opus 5.5: 1,293 (1,280–1,305) — claude-opus-5.5-high, high, 9 October 2026
  9. 9. Gemini 3 Pro: 1,289 (1,282–1,297) — gemini-3-pro, 9 October 2026
  10. 10. Gemini 3.8 Flash: 1,289 (1,279–1,300) — gemini-3.8-flash-high, high, 9 October 2026
  11. 11. Claude Opus 5: 1,287 (1,281–1,294) — claude-opus-5-high, high, 9 October 2026
  12. 12. GPT-5.6 Sol: 1,287 (1,280–1,294) — gpt-5.6-sol-xhigh, xhigh, 9 October 2026
  13. 13. Claude Fable 5.1: 1,286 (1,277–1,296) — claude-fable-5.1-max, max, 9 October 2026
  14. 14. GPT-6 Astra: 1,286 (1,275–1,297) — gpt-6-astra-max, max, 9 October 2026
  15. 15. GPT-5.5: 1,286 (1,280–1,292) — gpt-5.5, 9 October 2026
  16. 16. Claude Opus 4.8: 1,286 (1,279–1,292) — claude-opus-4-8-high, high, 9 October 2026
  17. 17. Gemini 3.5 Flash: 1,285 (1,278–1,292) — gemini-3.5-flash-medium, medium, 9 October 2026
  18. 18. GPT-5.4: 1,285 (1,279–1,291) — gpt-5.4-high, high, 9 October 2026
  19. 19. Gemini 3.6 Flash: 1,281 (1,273–1,289) — gemini-3.6-flash-high, high, 9 October 2026
  20. 20. Grok 4.5: 1,280 (1,273–1,287) — grok-4.5, 9 October 2026
  21. 21. Claude Sonnet 4.6: 1,275 (1,269–1,281) — claude-sonnet-4-6, 9 October 2026
  22. 22. Gemini 3 Flash: 1,273 (1,268–1,278) — gemini-3-flash, 9 October 2026
  23. 23. GPT-6 Sol: 1,270 (1,258–1,283) — gpt-6-sol-max, max, 9 October 2026
  24. 24. GPT-5.6 Terra: 1,268 (1,261–1,275) — gpt-5.6-terra-xhigh, xhigh, 9 October 2026
  25. 25. Claude Sonnet 5.5: 1,266 (1,252–1,280) — claude-sonnet-5.5-xhigh, xhigh, 9 October 2026
  26. 26. Gemini 3.5 Flash-Lite: 1,266 (1,258–1,274) — gemini-3.5-flash-lite, 9 October 2026
  27. 27. Claude Sonnet 5: 1,263 (1,256–1,270) — claude-sonnet-5-high, high, 9 October 2026
  28. 28. Grok 4.6: 1,262 (1,253–1,271) — grok-4.6-high, high, 9 October 2026
  29. 29. DeepSeek-V4.1-Flash: 1,260 (1,249–1,271) — deepseek-v4.1-flash-max, max, 9 October 2026
  30. 30. GPT-5.6 Luna: 1,258 (1,250–1,265) — gpt-5.6-luna-xhigh, xhigh, 9 October 2026
  31. 31. GPT-5.1: 1,251 (1,244–1,259) — gpt-5.1-high, high, 9 October 2026
  32. 32. GPT-5.4 Mini: 1,251 (1,245–1,257) — gpt-5.4-mini-high, high, 9 October 2026
  33. 33. Gemini 2.5 Pro: 1,248 (1,243–1,252) — gemini-2.5-pro, 9 October 2026
  34. 34. Grok 4.7: 1,248 (1,236–1,259) — grok-4.7-xhigh, xhigh, 9 October 2026
  35. 35. Qwen3.8-27B: 1,244 (1,236–1,253) — qwen3.8-27b, 9 October 2026
  36. 36. GPT-5.2: 1,243 (1,237–1,249) — gpt-5.2-high, high, 9 October 2026
  37. 37. Grok 4.3: 1,243 (1,237–1,248) — grok-4.3, 9 October 2026
  38. 38. GPT-6 Luna: 1,241 (1,229–1,253) — gpt-6-luna-max, max, 9 October 2026
  39. 39. Qwen3.5-122B-A10B: 1,226 (1,219–1,233) — qwen3.5-122b-a10b, 9 October 2026
  40. 40. GPT-5 Chat: 1,226 (1,219–1,233) — gpt-5-chat, 9 October 2026
  41. 41. GPT-4.5 Preview: 1,225 (1,213–1,237) — gpt-4.5-preview-2025-02-27, 9 October 2026
  42. 42. Qwen3.5-27B: 1,217 (1,211–1,222) — qwen3.5-27b, 9 October 2026
  43. 43. o3: 1,216 (1,209–1,222) — o3-2025-04-16, 9 October 2026
  44. 44. Gemini 2.5 Flash: 1,215 (1,211–1,220) — gemini-2.5-flash, 9 October 2026
  45. 45. GPT-4.1: 1,214 (1,207–1,221) — gpt-4.1-2025-04-14, 9 October 2026
  46. 46. GPT-5: 1,210 (1,203–1,217) — gpt-5-high, high, 9 October 2026
  47. 47. Claude Sonnet 4: 1,206 (1,189–1,223) — claude-sonnet-4-20250514-thinking-32k, thinking-32k, 9 October 2026
  48. 48. GPT-4.1 mini: 1,201 (1,193–1,208) — gpt-4.1-mini-2025-04-14, 9 October 2026
  49. 49. o4-mini: 1,201 (1,194–1,208) — o4-mini-2025-04-16, 9 October 2026
  50. 50. Mistral Medium 3.5: 1,200 (1,191–1,210) — mistral-medium-3.5, 9 October 2026
  51. 51. Mistral Large 3: 1,200 (1,192–1,208) — mistral-large-3, 9 October 2026
  52. 52. GPT-5.4 nano: 1,199 (1,193–1,206) — gpt-5.4-nano-high, high, 9 October 2026
  53. 53. Grok 4.1 Fast: 1,196 (1,189–1,203) — grok-4-1-fast-reasoning, reasoning, 9 October 2026
  54. 54. Claude Sonnet 3.7: 1,193 (1,177–1,208) — claude-3-7-sonnet-20250219-thinking-32k, thinking-32k, 9 October 2026
  55. 55. o1: 1,193 (1,182–1,203) — o1-2024-12-17, 9 October 2026
  56. 56. GPT-5 Mini: 1,183 (1,175–1,191) — gpt-5-mini-high, high, 9 October 2026
  57. 57. GPT-4o: 1,162 (1,153–1,170) — gpt-4o-2024-05-13, 9 October 2026
  58. 58. GPT-5 nano: 1,146 (1,135–1,158) — gpt-5-nano-high, high, 9 October 2026
  59. 59. Claude 3.5 Haiku: 1,127 (1,111–1,143) — claude-3-5-haiku-20241022, 9 October 2026
  60. 60. GPT-4 Turbo: 1,112 (1,100–1,123) — gpt-4-turbo-2024-04-09, 9 October 2026
  61. 61. GPT-4o mini: 1,097 (1,089–1,105) — gpt-4o-mini-2024-07-18, 9 October 2026
  62. 62. GPT-4.1 nano: 1,089 (1,070–1,107) — gpt-4.1-nano-2025-04-14, 9 October 2026
  63. 63. Qwen2-VL-72B: 1,084 (1,074–1,095) — qwen2-vl-72b, 9 October 2026
  64. 64. Claude 3 Opus: 1,061 (1,051–1,072) — claude-3-opus-20240229, 9 October 2026
  65. 65. Claude 3 Sonnet: 1,016 (1,005–1,027) — claude-3-sonnet-20240229, 9 October 2026
  66. 66. Claude 3 Haiku: 1,000 (987–1,012) — claude-3-haiku-20240307, 9 October 2026

Arena: the benchmark and its data · CC BY 4.0

Other benchmarks

All of the Model Tracker