Ir al contenido
AI.info

Model Tracker

OTIS Mock AIME 2024–2025 leaderboard

The models in the Model Tracker that Epoch AI has measured on OTIS Mock AIME 2024–2025, ranked by the highest score of each: the name Epoch AI gives it, the setting it ran and the day.

En inglés

The models in the Model Tracker that Epoch AI has measured on OTIS Mock AIME 2024–2025, 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. Gemini 4 Argon: 100.0% (± 0.0) — gemini-4-argon_high, high, 13 October 2026
  2. 1. GPT-6.1 Sol: 100.0% (± 0.0) — gpt-6.1-sol_max, max, 29 September 2026
  3. 1. Claude Sonnet 5.5: 100.0% (± 0.0) — claude-sonnet-5-5_max, max, 29 September 2026
  4. 1. GPT-6 Sol: 100.0% (± 0.0) — gpt-6-sol_max, max, 22 September 2026
  5. 1. Claude Opus 5.5: 100.0% (± 0.0) — claude-opus-5-5_max, max, 22 September 2026
  6. 1. Claude Fable 5.1: 100.0% (± 0.0) — claude-fable-5-1_max, max, 1 September 2026
  7. 1. GPT-6 Astra: 100.0% (± 0.0) — gpt-6-astra_max, max, 30 August 2026
  8. 1. GPT-5.6 Sol: 100.0% (± 0.0) — gpt-5.6-sol_max, max, 9 July 2026
  9. 1. Claude Fable 5: 100.0% (± 0.0) — claude-fable-5_high, high, 6 August 2026
  10. 1. GPT-5.5: 100.0% (± 0.0) — gpt-5.5-pre-release_xhigh, xhigh, 24 April 2026
  11. 1. GPT-5.5 Pro: 100.0% (± 0.0) — gpt-5.5-pro-pre-release_xhigh, xhigh, 24 April 2026
  12. 12. GPT-5.6 Terra: 99.7% (± 0.3) — gpt-5.6-terra_max, max, 9 July 2026
  13. 12. Claude Fable 5.2: 99.7% (± 0.3) — claude-fable-5-2_max, max, 6 November 2026
  14. 14. Muse Wonder: 99.4% (± 0.4) — muse-wonder-1.0_max, max, 12 October 2026
  15. 14. Qwen3.8-Max: 99.4% (± 0.4) — qwen3.8-max_xhigh, xhigh, 4 August 2026
  16. 14. Mistral Large 4: 99.4% (± 0.4) — mistral-large-4-0-oss_high, high, 2 November 2026
  17. 17. Grok 4.6: 99.2% (± 0.5) — grok-4.6_xhigh, xhigh, 14 August 2026
  18. 18. Claude Haiku 5.5: 98.9% (± 0.5) — claude-haiku-5-5_xhigh, xhigh, 7 October 2026
  19. 18. GPT-6 Luna: 98.9% (± 0.7) — gpt-6-luna_max, max, 22 September 2026
  20. 18. Gemini 3.8 Flash: 98.9% (± 0.9) — gemini-3.8-flash_high, high, 2 September 2026
  21. 18. Claude Opus 5: 98.9% (± 1.1) — claude-opus-5_max, max, 24 July 2026
  22. 22. DeepSeek-V4.1-Flash: 98.3% (± 0.6) — deepseek-v4.1-flash_max, max, 8 October 2026
  23. 22. GPT-5.6 Luna: 98.3% (± 1.2) — gpt-5.6-luna_max, max, 9 July 2026
  24. 22. Claude Opus 4.8: 98.3% (± 1.4) — claude-opus-4-8_max, max, 7 June 2026
  25. 25. Grok 4.7: 98.1% (± 1.4) — grok-4.7_xhigh, xhigh, 22 September 2026
  26. 26. Claude Opus 4.7: 97.8% (± 2.2) — claude-opus-4-7_xhigh, xhigh, 17 April 2026
  27. 27. Grok 4.5: 97.8% (± 1.3) — grok-4.5_high, high, 8 July 2026
  28. 27. GPT-5.4: 97.8% (± 2.2) — gpt-5.4-2026-03-05_high, high, 15 July 2026
  29. 29. Gemini 3.7 Flash: 97.2% (± 1.9) — gemini-3.7-flash_high, high, 14 August 2026
  30. 30. DeepSeek V4 Pro: 96.7% (± 2.0) — deepseek-v4-pro_max, max, 17 June 2026
  31. 31. GPT-5.2: 96.1% (± 2.6) — gpt-5.2-2025-12-11_high, high, 11 December 2025
  32. 32. Gemini 3.1 Pro: 95.6% (± 3.1) — gemini-3.1-pro-preview, 20 February 2026
  33. 33. Gemini 3.5 Flash: 95.6% (± 2.7) — gemini-3.5-flash_high, high, 25 May 2026
  34. 33. Gemini 3 Flash: 95.6% (± 3.1) — gemini-3-flash-preview_high, high, 6 August 2026
  35. 35. Claude Sonnet 5: 94.7% (± 2.3) — claude-sonnet-5_xhigh, xhigh, 1 July 2026
  36. 36. Claude Opus 4.6: 94.4% (± 2.8) — claude-opus-4-6_64K, 64K, 6 February 2026
  37. 37. Gemini 3.6 Flash: 94.2% (± 3.1) — gemini-3.6-flash_high, high, 2 August 2026
  38. 38. Grok 4.20: 92.2% (± 3.3) — grok-4.20-0309-reasoning, 13 July 2026
  39. 39. GPT-5: 91.4% (± 3.7) — gpt-5-2025-08-07_high, high, 29 October 2025
  40. 39. Gemini 3 Pro: 91.4% (± 3.7) — gemini-3-pro-preview, 19 November 2025
  41. 41. Qwen3.6-27B: 91.1% (± 4.3) — qwen3.6-27b, 7 August 2026
  42. 42. Muse Spark: 88.9% (± 4.7) — muse-spark, 8 April 2026
  43. 43. gpt-oss-120b: 88.9% (± 4.5) — openai/gpt-oss-120b_high, high, 11 December 2025
  44. 43. GPT-5.4 Mini: 88.9% (± 4.7) — gpt-5.4-mini-2026-03-17_xhigh, xhigh, 7 August 2026
  45. 45. GPT-5.1: 88.6% (± 4.0) — gpt-5.1-2025-11-13_high, high, 13 November 2025
  46. 46. GPT-5.4 nano: 87.8% (± 3.9) — gpt-5.4-nano-2026-03-17_high, high, 14 April 2026
  47. 47. Qwen3.6-35B-A3B: 86.7% (± 5.1) — qwen3.6-35b-a3b, 7 August 2026
  48. 47. GPT-5 Mini: 86.7% (± 4.3) — gpt-5-mini-2025-08-07_high, high, 30 October 2025
  49. 49. Claude Opus 4.5: 86.1% (± 4.3) — claude-opus-4-5-20251101_32K, 32K, 24 November 2025
  50. 50. Claude Sonnet 4.6: 85.8% (± 4.5) — claude-sonnet-4-6_32K, 32K, 20 February 2026
  51. 51. Gemini 2.5 Pro: 84.7% (± 4.8) — gemini-2.5-pro, 16 November 2025
  52. 52. o3: 84.4% (± 5.5) — o3-2025-04-16_medium, medium, 7 August 2026
  53. 53. o4-mini: 81.7% (± 4.7) — o4-mini-2025-04-16_high, high, 16 April 2025
  54. 54. GPT-5 nano: 81.1% (± 4.6) — gpt-5-nano-2025-08-07_high, high, 31 October 2025
  55. 55. Gemini 3.1 Flash-Lite: 80.0% (± 6.0) — gemini-3.1-flash-lite_high, high, 6 August 2026
  56. 56. Claude Sonnet 4.5: 77.8% (± 6.3) — claude-sonnet-4-5-20250929_59K, 59K, 28 October 2025
  57. 56. Grok 3 mini: 77.8% (± 6.3) — grok-3-mini-beta_high, high, 10 April 2025
  58. 58. o3-mini: 76.9% (± 5.2) — o3-mini-2025-01-31_high, high, 27 February 2025
  59. 59. o1: 73.3% (± 6.7) — o1-2024-12-17_high, high, 15 July 2026
  60. 60. Gemini 2.5 Flash: 73.1% (± 5.5) — gemini-2.5-flash-preview-04-17, 22 April 2025
  61. 61. Gemini 3.5 Flash-Lite: 71.1% (± 6.8) — gemini-3.5-flash-lite_high, high, 6 August 2026
  62. 61. Claude Sonnet 4: 71.1% (± 6.8) — claude-sonnet-4-20250514_32K, 32K, 22 May 2025
  63. 63. Qwen3.5-35B-A3B: 70.0% (± 6.0) — qwen3.5-35b-a3b_none, none, 30 August 2026
  64. 64. Claude Opus 4.1: 68.9% (± 7.0) — claude-opus-4-1-20250805_27K, 27K, 5 August 2025
  65. 65. Qwen3-32B: 66.9% (± 6.0) — Qwen3-32B, 30 August 2026
  66. 66. Claude Haiku 4.5: 66.7% (± 7.1) — claude-haiku-4-5-20251001_32K, 32K, 22 October 2025
  67. 67. Qwen3-14B: 66.4% (± 6.2) — qwen3-14b, 28 August 2026
  68. 67. DeepSeek-R1: 66.4% (± 4.9) — DeepSeek-R1-0528, 29 May 2025
  69. 69. gpt-oss-20b: 65.3% (± 5.9) — gpt-oss-20b_medium, medium, 27 August 2026
  70. 70. Qwen3-30B-A3B: 62.8% (± 6.2) — Qwen3-30B-A3B, 28 August 2026
  71. 71. Qwen3.5-9B: 61.7% (± 6.0) — qwen3.5-9B_none, none, 27 August 2026
  72. 72. Claude Sonnet 3.7: 57.8% (± 7.4) — claude-3-7-sonnet-20250219_64K, 64K, 13 March 2025
  73. 73. Qwen3-8B: 56.1% (± 6.5) — qwen3-8b, 27 August 2026
  74. 74. Qwen3.5-4B: 55.8% (± 6.2) — qwen3.5-4b_none, none, 28 August 2026
  75. 75. Grok 3: 55.6% (± 7.5) — grok-3-beta, 10 April 2025
  76. 76. o1-mini: 46.9% (± 6.5) — o1-mini-2024-09-12_high, high, 6 March 2025
  77. 77. GPT-4.1 mini: 44.7% (± 6.1) — gpt-4.1-mini-2025-04-14, 14 April 2025
  78. 78. GPT-4.1: 38.3% (± 6.3) — gpt-4.1-2025-04-14, 14 April 2025
  79. 79. GPT-4.5: 37.8% (± 7.3) — gpt-4.5-preview-2025-02-27, 28 February 2025
  80. 79. DeepSeek-V3: 37.8% (± 6.2) — DeepSeek-V3-0324, 1 April 2025
  81. 81. Mistral Medium 3: 32.2% (± 6.0) — mistral-medium-2505, 7 May 2025
  82. 82. o1-preview: 31.1% (± 7.0) — o1-preview-2024-09-12, 7 March 2025
  83. 82. Gemini 2.0 Flash: 31.1% (± 6.3) — gemini-2.0-flash-001, 25 February 2025
  84. 84. Mistral Small 3.2: 30.3% (± 5.8) — mistral-small-3.2-2506, 30 August 2026
  85. 85. Magistral Small 1.0: 30.0% (± 5.9) — magistral-small-2506, 6 August 2026
  86. 86. GPT-4.1 nano: 28.9% (± 6.0) — gpt-4.1-nano-2025-04-14, 14 April 2025
  87. 87. Magistral Small 1.2: 28.1% (± 5.6) — magistral-small-2509, 30 August 2026
  88. 88. Gemini 1.5 Pro: 23.1% (± 5.2) — gemini-1.5-pro-002, 25 February 2025
  89. 89. Llama 4 Maverick: 20.6% (± 4.9) — Llama-4-Maverick-17B-128E-Instruct-FP8, 8 April 2025
  90. 90. Gemini 1.5 Flash: 16.3% (± 4.1) — gemini-1.5-flash-002, 25 February 2025
  91. 91. Grok-2: 11.5% (± 3.4) — grok-2-1212, 25 February 2025
  92. 92. Qwen3-1.7B: 8.1% (± 2.8) — qwen3-1.7b_none, none, 28 August 2026
  93. 93. Llama 4 Scout: 7.8% (± 3.0) — Llama-4-Scout-17B-16E-Instruct, 8 April 2025
  94. 94. GPT-4o mini: 6.9% (± 2.7) — gpt-4o-mini-2024-07-18, 30 July 2025
  95. 95. GPT-4 Turbo: 6.7% (± 2.4) — gpt-4-turbo-2024-04-09, 27 February 2025
  96. 96. GPT-4o: 6.4% (± 2.6) — gpt-4o-2024-08-06, 25 February 2025
  97. 97. Mistral Small 3.1: 5.8% (± 2.3) — mistral-small-2503, 18 March 2025
  98. 98. Claude 3 Opus: 4.7% (± 2.0) — claude-3-opus-20240229, 25 February 2025
  99. 99. Gemini 1.5 Flash-8B: 4.6% (± 2.0) — gemini-1.5-flash-8b-001, 25 February 2025
  100. 100. Claude 3.5 Haiku: 4.3% (± 1.6) — claude-3-5-haiku-20241022, 25 February 2025
  101. 101. Claude 2: 2.5% (± 1.5) — claude-2.0, 7 March 2025
  102. 101. Claude 3 Sonnet: 2.5% (± 2.0) — claude-3-sonnet-20240229, 25 February 2025
  103. 103. GPT-3.5 Turbo: 2.2% (± 2.2) — gpt-3.5-turbo-0125, 7 August 2026
  104. 104. Claude 2.1: 1.9% (± 1.0) — claude-2.1, 7 March 2025
  105. 105. Claude 3 Haiku: 1.8% (± 0.8) — claude-3-haiku-20240307, 25 February 2025
  106. 106. GPT-4: 1.1% (± 0.5) — gpt-4-0613, 23 October 2025

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

Other benchmarks

All of the Model Tracker