Vai al contenuto
AI.info

Model Tracker

Epoch Capabilities Index leaderboard

The models in the Model Tracker that Epoch AI has measured on Epoch Capabilities Index, ranked by the highest score of each: the name Epoch AI gives it, the setting it ran and the day. Epoch AI, ‘Capabilities & benchmarking’, epoch.ai/bench

In inglese

The models in the Model Tracker that Epoch AI has measured on Epoch Capabilities Index, ranked by the highest score of each: the name Epoch AI gives it, the setting it ran and the day. Epoch AI, ‘Capabilities & benchmarking’, epoch.ai/benchmarks, CC BY 4.0: Epoch AI’s own runs and its Capabilities Index.

  1. 1. Claude Opus 5.5: 166.8 (163.5–171.0) — Claude Opus 5.5, 19 October 2026
  2. 2. GPT-6 Astra: 166.4 (163.4–171.2) — GPT-6 Astra, 19 October 2026
  3. 3. GPT-6.1 Sol: 166.3 (163.2–170.8) — GPT-6.1 Sol, 19 October 2026
  4. 4. Claude Fable 5.1: 164.7 (161.9–168.5) — Claude Fable 5.1, 19 October 2026
  5. 5. Gemini 4 Argon: 164.6 (161.7–168.7) — Gemini 4 Argon, 19 October 2026
  6. 6. Claude Sonnet 5.5: 164.5 (161.4–168.3) — Claude Sonnet 5.5, 19 October 2026
  7. 7. Claude Opus 5: 162.6 (160.0–165.8) — Claude Opus 5, 19 October 2026
  8. 8. GPT-6 Sol: 162.6 (160.0–165.8) — GPT-6 Sol, 19 October 2026
  9. 9. Claude Fable 5: 162.1 (159.5–166.3) — Claude Fable 5, 19 October 2026
  10. 10. GPT-5.5 Pro: 162.0 (159.0–165.1) — GPT-5.5 Pro, 19 October 2026
  11. 11. GPT-5.6 Sol: 161.7 (159.5–165.0) — GPT-5.6 Sol, 19 October 2026
  12. 12. GPT-5.6 Terra: 159.4 (156.8–162.2) — GPT-5.6 Terra, 19 October 2026
  13. 13. GPT-5.4 Pro: 159.0 (156.5–161.7) — GPT-5.4 Pro, 19 October 2026
  14. 14. Claude Haiku 5.5: 159.0 (156.6–162.1) — Claude Haiku 5.5, 19 October 2026
  15. 15. GPT-5.5: 159.0 (156.4–161.9) — GPT-5.5, 19 October 2026
  16. 16. Claude Opus 4.8: 158.3 (156.2–160.8) — Claude Opus 4.8, 19 October 2026
  17. 17. Gemini 3.7 Flash: 157.2 (155.3–159.7) — Gemini 3.7 Flash, 19 October 2026
  18. 18. Gemini 3.8 Flash: 157.2 (155.0–160.5) — Gemini 3.8 Flash, 19 October 2026
  19. 19. Qwen3.8-Max: 156.9 (155.1–159.1) — Qwen 3.8 Max, 19 October 2026
  20. 20. GPT-5.3-Codex: 156.8 (152.9–160.5) — GPT-5.3 Codex, 19 October 2026
  21. 21. GPT-5.4: 156.7 (155.0–158.9) — GPT-5.4, 19 October 2026
  22. 22. Claude Sonnet 5: 156.4 (154.0–158.9) — Claude Sonnet 5, 19 October 2026
  23. 23. GPT-5.6 Luna: 156.3 (153.9–159.0) — GPT-5.6 Luna, 19 October 2026
  24. 24. Claude Opus 4.7: 156.3 (154.6–158.5) — Claude Opus 4.7, 19 October 2026
  25. 25. GPT-6 Luna: 156.3 (153.9–158.8) — GPT-6 Luna, 19 October 2026
  26. 25. Grok 4.6: 156.3 (154.5–158.7) — Grok 4.6, 19 October 2026
  27. 27. GPT-5.2 Pro: 156.0 (153.8–159.6) — GPT-5.2 Pro, 19 October 2026
  28. 28. DeepSeek-V4.1-Flash: 155.6 (153.5–157.6) — DeepSeek V4.1 Flash, 19 October 2026
  29. 29. Claude Opus 4.6: 155.3 (153.6–157.1) — Claude Opus 4.6, 19 October 2026
  30. 30. Gemini 3.1 Pro: 154.7 (152.5–157.1) — Gemini 3.1 Pro, 19 October 2026
  31. 31. Gemini 3.5 Flash: 154.4 (152.4–156.5) — Gemini 3.5 Flash, 19 October 2026
  32. 32. Grok 4.7: 154.4 (152.7–156.5) — Grok 4.7, 19 October 2026
  33. 33. Gemini 3.6 Flash: 154.2 (152.7–156.1) — Gemini 3.6 Flash, 19 October 2026
  34. 34. Grok 4.5: 153.9 (152.5–155.9) — Grok 4.5, 19 October 2026
  35. 35. GPT-5.2: 153.5 (151.6–155.3) — GPT-5.2, 19 October 2026
  36. 36. Gemini 3 Pro: 152.9 (151.0–155.0) — Gemini 3 Pro, 19 October 2026
  37. 37. Claude Sonnet 4.6: 152.4 (150.0–154.7) — Claude Sonnet 4.6, 19 October 2026
  38. 38. Muse Spark: 152.1 (150.3–155.4) — Muse Spark, 19 October 2026
  39. 39. Grok 4.20: 152.1 (149.5–154.2) — Grok 4.20, 19 October 2026
  40. 40. Gemini 3 Flash: 151.8 (150.5–153.4) — Gemini 3 Flash, 19 October 2026
  41. 41. Qwen3.8-27B: 151.6 (149.8–153.1) — Qwen 3.8 27B, 19 October 2026
  42. 42. GPT-5 Pro: 150.7 (148.9–154.0) — GPT-5 Pro, 19 October 2026
  43. 43. Claude Opus 4.5: 150.2 (148.0–152.8) — Claude Opus 4.5, 19 October 2026
  44. 44. GPT-5: 150.0 — GPT-5, 19 October 2026
  45. 45. GPT-5.1: 149.8 (148.5–151.5) — GPT-5.1, 19 October 2026
  46. 46. DeepSeek V4 Pro: 149.2 (147.4–150.7) — DeepSeek-V4-Pro, 19 October 2026
  47. 47. GPT-5.4 Mini: 148.8 (147.1–150.5) — GPT-5.4 Mini, 19 October 2026
  48. 48. o3-pro: 147.4 (145.7–150.2) — o3-pro, 19 October 2026
  49. 49. Mistral Large 4: 146.9 (143.7–150.9) — Mistral Large 4, 19 October 2026
  50. 50. o3: 146.9 (145.0–148.7) — o3, 19 October 2026
  51. 51. Claude Sonnet 4.5: 146.8 (145.2–148.4) — Claude Sonnet 4.5, 19 October 2026
  52. 52. Qwen3.6-27B: 146.5 (144.1–147.8) — Qwen3.6 27B, 19 October 2026
  53. 53. Grok 4: 146.4 (144.6–148.2) — Grok 4, 19 October 2026
  54. 54. GPT-5.4 nano: 146.1 (143.6–148.5) — GPT-5.4 Nano, 19 October 2026
  55. 55. DeepSeek-V4-Flash: 146.1 (143.7–148.1) — DeepSeek-V4-Flash, 19 October 2026
  56. 56. o4-mini: 145.6 (143.3–147.3) — o4-mini, 19 October 2026
  57. 57. GPT-5 Mini: 145.4 (143.5–146.9) — GPT-5 mini, 19 October 2026
  58. 58. Gemini 2.5 Pro: 145.3 (143.5–146.9) — Gemini 2.5 Pro (Jun 2025), 19 October 2026
  59. 59. Gemini 3.5 Flash-Lite: 145.2 (142.7–146.7) — Gemini 3.5 Flash-Lite, 19 October 2026
  60. 60. DeepSeek-V3.2-Exp: 145.0 (142.4–147.0) — DeepSeek-V3.2-Exp, 19 October 2026
  61. 61. Gemini 3.1 Flash-Lite: 144.5 (142.3–146.0) — Gemini 3.1 Flash-Lite, 19 October 2026
  62. 62. Grok 4 Fast: 144.2 (142.0–146.4) — Grok 4 Fast, 19 October 2026
  63. 63. Qwen3.6-35B-A3B: 144.1 (141.5–146.3) — Qwen 3.6 35B-A3B, 19 October 2026
  64. 64. Claude Opus 4.1: 144.1 (142.0–146.1) — Claude Opus 4.1, 19 October 2026
  65. 65. Gemini 2.5 Flash: 143.0 (139.5–146.6) — Gemini 2.5 Flash (Sep 2025), 19 October 2026
  66. 66. Qwen3.5-35B-A3B: 142.5 (140.1–144.9) — Qwen3.5-35B-A3B, 19 October 2026
  67. 67. Claude Haiku 4.5: 142.4 (139.3–144.1) — Claude Haiku 4.5, 19 October 2026
  68. 68. o1: 141.8 (139.5–142.9) — o1, 19 October 2026
  69. 69. Claude Sonnet 4: 141.7 (138.7–143.1) — Claude Sonnet 4, 19 October 2026
  70. 70. Mistral Medium 3.5: 141.5 (138.7–144.0) — Mistral Medium 3.5, 19 October 2026
  71. 71. DeepSeek-R1: 141.3 (138.7–142.8) — DeepSeek-R1 (May 2025), 19 October 2026
  72. 72. Claude Sonnet 3.7: 141.1 (138.5–142.8) — Claude 3.7 Sonnet, 19 October 2026
  73. 73. o3-mini: 140.3 (137.2–141.9) — o3-mini, 19 October 2026
  74. 73. Grok 3 mini: 140.3 (137.6–142.2) — Grok-3 mini, 19 October 2026
  75. 75. gpt-oss-120b: 140.0 (135.9–142.4) — gpt-oss-120b, 19 October 2026
  76. 76. DeepSeek-V3.1: 139.9 (135.8–143.2) — DeepSeek-V3.1, 19 October 2026
  77. 77. Qwen3-235B-A22B: 139.4 (135.2–140.8) — Qwen3-235B-A22B, 19 October 2026
  78. 78. GPT-5 nano: 139.3 (134.8–141.4) — GPT-5 nano, 19 October 2026
  79. 79. Qwen3-32B: 138.5 (134.8–140.4) — Qwen3-32B, 19 October 2026
  80. 80. Grok 3: 138.3 (135.5–139.7) — Grok 3, 19 October 2026
  81. 81. Qwen3-14B: 138.2 (133.8–140.0) — Qwen3-14B, 19 October 2026
  82. 82. gpt-oss-20b: 137.8 (133.2–139.8) — gpt-oss-20b, 19 October 2026
  83. 83. GPT-4.1: 136.8 (133.7–138.6) — GPT-4.1, 19 October 2026
  84. 84. GPT-4.5: 136.7 (133.4–138.7) — GPT-4.5, 19 October 2026
  85. 85. Qwen3-30B-A3B: 136.2 (130.7–138.7) — Qwen3-30B-A3B, 19 October 2026
  86. 86. Qwen3-8B: 136.2 (130.8–138.3) — Qwen3-8B, 19 October 2026
  87. 87. DeepSeek-V3: 135.9 (132.5–137.8) — DeepSeek-V3 (Mar 2025), 19 October 2026
  88. 88. o1-mini: 135.8 (132.0–137.3) — o1-mini, 19 October 2026
  89. 89. GPT-4.1 mini: 135.0 (131.2–136.7) — GPT-4.1 mini, 19 October 2026
  90. 90. o1-preview: 134.8 (131.1–138.4) — o1-preview, 19 October 2026
  91. 91. Gemini 2.0 Flash: 134.7 (124.3–136.6) — Gemini 2.0 Flash (Dec 2024), 19 October 2026
  92. 92. Mistral Medium 3: 134.1 (129.9–135.5) — Mistral Medium 3, 19 October 2026
  93. 93. Gemini 2.5 Flash-Lite: 134.0 (130.3–136.5) — Gemini 2.5 Flash-Lite (Jun 2025), 19 October 2026
  94. 94. Magistral Small 1.0: 133.2 (129.0–134.4) — Magistral Small 1.0, 19 October 2026
  95. 95. Llama 4 Maverick: 132.2 (127.7–134.1) — Llama 4 Maverick, 19 October 2026
  96. 96. Mistral Small 3.2: 131.8 (126.8–133.9) — Mistral Small 3.2, 19 October 2026
  97. 97. Gemini 1.5 Pro: 131.7 (127.1–133.1) — Gemini 1.5 Pro (Sept 2024), 19 October 2026
  98. 98. Magistral Small 1.2: 131.4 (125.9–133.4) — Magistral Small 1.2, 19 October 2026
  99. 99. Grok-2: 130.5 (125.6–132.3) — Grok-2 (Dec 2024), 19 October 2026
  100. 100. Llama 4 Scout: 129.7 (124.7–131.3) — Llama 4 Scout, 19 October 2026
  101. 101. GPT-4.1 nano: 129.6 (123.5–132.1) — GPT-4.1 nano, 19 October 2026
  102. 102. Gemini 1.5 Flash: 129.4 (123.1–131.3) — Gemini 1.5 Flash (Sep 2024), 19 October 2026
  103. 103. GPT-4o: 129.1 (123.7–132.0) — GPT-4o (May 2024), 19 October 2026
  104. 104. Mistral Small 3.1: 127.5 (122.1–129.5) — Mistral Small 3.1, 19 October 2026
  105. 105. GPT-4 Turbo: 127.3 (121.6–128.8) — GPT-4 Turbo (Apr 2024), 19 October 2026
  106. 106. Claude 3.5 Haiku: 127.2 (120.1–129.7) — Claude 3.5 Haiku, 19 October 2026
  107. 107. Claude 3 Opus: 126.9 (121.8–129.8) — Claude 3 Opus, 19 October 2026
  108. 108. GPT-4o mini: 126.6 (119.5–128.6) — GPT-4o mini, 19 October 2026
  109. 109. GPT-4: 125.9 (119.9–130.6) — GPT-4 (Mar 2023), 19 October 2026
  110. 110. Mixtral 8x22B: 122.1 (115.2–124.6) — Mixtral 8x22B, 19 October 2026
  111. 111. Claude 3 Sonnet: 120.7 (112.5–123.9) — Claude 3 Sonnet, 19 October 2026
  112. 112. Claude 2: 120.2 (113.7–123.9) — Claude 2, 19 October 2026
  113. 113. Claude 2.1: 119.3 (112.2–121.9) — Claude 2.1, 19 October 2026
  114. 114. GPT-3.5 Turbo: 118.6 (109.7–121.8) — GPT-3.5 Turbo (Nov 2023), 19 October 2026
  115. 115. Mixtral 8x7B: 118.5 (110.9–121.2) — Mixtral 8x7B, 19 October 2026
  116. 116. Claude 3 Haiku: 118.4 (110.4–121.7) — Claude 3 Haiku, 19 October 2026
  117. 117. Qwen-14B: 113.1 (104.6–117.7) — Qwen-14B, 19 October 2026
  118. 118. Qwen-7B: 106.9 (96.1–112.3) — Qwen-7B, 19 October 2026

Epoch AI: the benchmark and its data · CC BY 4.0

Other benchmarks

All of the Model Tracker