Benchmark Almanac

Pedagogy Benchmark

Gemini 4 Argon 92.9% leads since September 2026
Not in model cards

Models answer multiple-choice questions about teaching strategies and assessment from Chilean teacher exams.

Fab AI publishes overall accuracy for 283 model configurations. The paper scores 899 questions and uses 21 more as examples. The dataset requires acceptance of access conditions.

Made by
Fab AI
Released
December 2024
Kind
Exam questions
Level
University
Scored
Checked automatically
Answers
Multiple choice
Tasks
All public
Measures
Capability
Subject
Education
Language
English
In and out
Text → Text
Interaction
Single turn
Run by
Anyone
Occupations
Educational instruction and library
Strongest model tested
GPT-6 Astra #4 on the capability index
Last model added
October 2026
Scores refreshed
November 20th, 2026
259
Entries on its own leaderboard, 216 models we track

Gemini 4 Argon leads with 92.9% Gemini 4 Argon leads with 92.9%

Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.

259 entries. Scroll sideways for the rest.

  1. 92.9 Gemini 4 Argon
  2. 92.7 GPT-6 Astra
  3. 92.5 GPT-6.1 Terra
  4. 92.2 Claude Fable 5.1
  5. 92.1 GPT-5.5
  6. 91.9 Gemini 3.8 Flash
  7. 91.9 Claude Opus 5.5
  8. 91.8 GPT-6.1 Sol
  9. 91.5 Gemini 3.1 Pro
  10. 91.5 Grok 4.6
  11. 91.2 GPT-5.6 Sol
  12. 91.1 Claude Opus 5
  13. 90.9 Gemini 3.6 Flash
  14. 90.8 Gemini 3.7 Flash
  15. 90.5 Gemini 3 Pro
  16. 90.5 Claude Fable 5
  17. 90.3 Qwen3.8 Max (0902)
  18. 90.1 Qwen3.7-Max
  19. 90.1 Muse Spark 1.1
  20. 89.7 Claude Sonnet 4.6 Adaptive Thinking
  21. 89.7 Qwen 3.6 Max (Preview)
  22. 89.7 Grok 4.5
  23. 89.7 Muse Wonder
  24. 89.4 Gemini 3 Flash
  25. 89.4 Kimi K2.6
  26. 89.4 Muse Spark 1.2
  27. 89.4 Grok 4.7
  28. 89.3 Claude Sonnet 4.5
  29. 89.3 Tencent Hy4 preview
  30. 89.2 Claude Opus 4.8
  31. 89.2 Qwen 3.8 Max
  32. 89.1 DeepSeek-V4-Pro
  33. 89.0 Claude Opus 4.7
  34. 89.0 Muse Spark 1.3
  35. 88.9 Qwen3.5-Plus
  36. 88.8 Claude Opus 4.7 Adaptive Thinking
  37. 88.8 Gemini 2.5 Pro
  38. 88.8 Qwen3.8 Flash
  39. 88.7 GPT-5.4
  40. 88.7 Grok 4.3
  41. 88.4 Grok 4
  42. 88.4 Gemini 3.5 Flash
  43. 88.4 Kimi K3.1 Pro
  44. 88.3 Claude Sonnet 5.5
  45. 88.2 Claude Sonnet 5
  46. 88.1 Qwen3.5 397B-A17B
  47. 88.1 Qwen 3.6 Plus
  48. 88.0 GLM-5.1
  49. 88.0 Qwen3.8-2.4T-A95B
  50. 87.9 o3
  51. 87.9 Claude Opus 4.6
  52. 87.8 GPT-5
  53. 87.8 GPT-5.1
  54. 87.8 Kimi K2.5
  55. 87.8 Kimi K3
  56. 87.8 DeepSeek V4.1 Flash
  57. 87.7 GLM-5
  58. 87.7 Qwen3.5-27B
  59. 87.7 Inkling
  60. 87.5 GPT-5.2
  61. 87.5 Grok 4.20
  62. 87.5 Qwen3.6 27B
  63. 87.4 Claude Opus 4 Thinking (low)
  64. 87.4 Claude Sonnet 4.6
  65. 87.3 Qwen3.5-122B-A10B
  66. 87.2 GPT-5.6 Terra
  67. 87.2 DeepSeek V4.1 Flash 1017
  68. 87.0 Gemma 4 31B IT
  69. 87.0 DeepSeek-V4-Flash
  70. 87.0 GLM-5.3
  71. 86.9 Claude Opus 4.5
  72. 86.8 Claude Sonnet 4 Thinking (low)
  73. 86.8 Qwen 3.8 27B
  74. 86.7 Deepseek R1 (May '25)
  75. 86.5 Qwen3.5-Flash
  76. 86.4 o1
  77. 86.3 Claude Opus 4
  78. 86.3 GLM-4.5
  79. 86.3 Grok 4 Fast
  80. 86.3 Gemini 2.5 Flash
  81. 86.2 Nemotron 3 Ultra
  82. 85.9 GLM-5.2
  83. 85.9 Claude Haiku 5.5
  84. 85.7 Qwen3-235B-A22B-Thinking
  85. 85.4 Claude-3.7 Sonnet Thinking (medium)
  86. 85.3 Doubao-Seed-1.6
  87. 85.3 Gemini 3.1 Flash-Lite
  88. 85.3 MiMo-V2.6 Pro
  89. 85.2 GPT-4.5
  90. 85.2 Mistral Large 4
  91. 85.1 Grok 4.1 Fast
  92. 85.1 Qwen 3.6 35B-A3B
  93. 85.1 Inkling-Small
  94. 84.9 Kimi K2 Thinking
  95. 84.8 Claude Sonnet 4
  96. 84.8 Qwen3.7 Flash
  97. 84.7 GLM-4.7
  98. 84.7 GPT-5.4 Mini
  99. 84.7 MiniMax-M3
  100. 84.5 Qwen3.5-35B-A3B
  101. 84.1 MiMo-V2-Pro
  102. 83.9 MiMo-V2.5-Pro
  103. 83.9 Qwen 3.6 Flash
  104. 83.9 Muse Glimmer
  105. 83.3 GPT-5 mini
  106. 83.2 Gemini 3.5 Flash-Lite
  107. 83.0 MiMo-V2.6 Flash
  108. 82.9 GLM-4.6
  109. 82.9 Gemma 4 26B A4B
  110. 82.9 MiMo-V2.5
  111. 82.8 MiniMax-M2.7
  112. 82.4 Qwen3-235B-A22B
  113. 82.4 Qwen3-32B
  114. 82.3 GPT-4.1
  115. 82.3 GPT-5.6 Luna
  116. 82.2 Mistral Medium 3.1
  117. 82.2 Kimi K2
  118. 82.0 o4-mini
  119. 81.8 Grok 3
  120. 81.8 Mistral Medium 3.5
  121. 81.7 Llama 4 Maverick
  122. 81.7 Gemma 4 12B
  123. 81.4 Qwen3.5-9B
  124. 81.2 Mistral Large 3
  125. 81.0 MiniMax-M3.1-Flash
  126. 80.8 Doubao-1.5-pro
  127. 80.7 Grok-3 mini
  128. 80.2 Grok-2
  129. 80.0 Mistral Medium 3
  130. 79.9 Gemini 2.0 Flash
  131. 79.8 Qwen3-30B-A3B
  132. 79.5 Claude Haiku 4.5
  133. 79.5 DeepSeek-V3.2
  134. 79.4 o3-mini
  135. 79.1 MiniMax-M2.1
  136. 78.8 Llama 3.1-405B
  137. 78.8 Nemotron 3.5 Lightning 30B-A3B
  138. 78.5 Qwen3-14B
  139. 78.4 Seed 1.6 Flash
  140. 78.3 DeepSeek-V3
  141. 78.3 GPT-4o
  142. 78.0 DeepSeek-V3.1
  143. 77.9 Command A
  144. 77.6 Gemini 2.0 Flash-Lite
  145. 77.6 Gemini 2.5 Flash-Lite
  146. 77.3 LearnLM 1.5 Pro
  147. 77.0 Claude 3 Opus
  148. 76.9 Qwen2.5-32B
  149. 76.8 Claude-3.5 Sonnet October
  150. 76.8 Claude 3.7 Sonnet
  151. 76.3 Llama 3.2 90B
  152. 76.0 Olmo 3.1 32B Think
  153. 75.9 Mistral Small 3.1 24B
  154. 75.9 Command A+
  155. 75.6 Qwen2.5-14B
  156. 75.6 GPT-5.4 Nano
  157. 75.4 Qwen2.5-72B
  158. 75.4 Mistral Small 3.2
  159. 75.3 Mistral Large November
  160. 75.3 Nova Pro
  161. 75.2 Llama 3.3 70B
  162. 75.2 gpt-oss-120b
  163. 75.1 GPT-4.1 mini
  164. 75.1 Magistral Medium 1.0
  165. 75.0 Llama 3.1-70B
  166. 74.9 GLM-4.5-Air
  167. 74.5 Llama 4 Scout
  168. 74.4 Gemma 4 E4B
  169. 74.3 GPT-4 Turbo
  170. 74.1 GPT-4
  171. 74.0 glm-4.7-flash
  172. 73.9 Gemini 1.5 Pro
  173. 73.9 gpt-oss-20b
  174. 73.8 Nemotron 3 Nano 30B-A3B
  175. 73.5 Qwen3-8B
  176. 73.1 Claude-3.5 Sonnet June
  177. 72.9 Grok Beta
  178. 72.6 o1-mini
  179. 72.6 Gemma 3 27B
  180. 72.6 Mistral Small 4
  181. 72.4 GPT-5 nano
  182. 72.2 Gemini 1.5 Flash
  183. 72.2 Phi-4
  184. 72.1 Doubao-1.5 Lite
  185. 72.0 LFM2.5-24B-A2B
  186. 71.8 Yi-Lightning
  187. 71.5 Yi-Large
  188. 71.5 Ministral 3 14B
  189. 71.4 Jamba Large 1.7
  190. 71.2 Gemma 2 27B
  191. 71.1 Granite 4.2 8B
  192. 70.8 sonar
  193. 70.4 Mistral Small 3
  194. 70.4 Qwen3-4B
  195. 70.0 Nova Lite
  196. 69.9 Jamba Large 1.6
  197. 69.6 Nova 2 Lite
  198. 69.4 Ministral 3 8B
  199. 69.2 GPT-4o mini
  200. 69.1 Jamba Large 1.5
  201. 68.2 Hunyuan Large
  202. 67.6 Gemma 3 12B
  203. 67.6 MiMo-V2-Flash
  204. 67.5 Magistral Small
  205. 66.7 Qwen2.5-7B
  206. 66.4 Olmo-3 7B Think
  207. 65.8 Mixtral 8x22B
  208. 65.4 Hunyuan Large (Long Context)
  209. 65.4 Olmo-3.1 32B
  210. 64.0 Gemma 3n E4B
  211. 63.5 GPT-4.1 nano
  212. 63.3 Claude 3.5 Haiku
  213. 63.2 Nova Micro
  214. 62.7 LFM3-1.7B
  215. 62.6 Gemma 4 E2B
  216. 61.9 LFM-7B
  217. 61.5 Command-R+ August
  218. 61.3 Ministral 3 3B
  219. 61.2 Jamba Instruct
  220. 61.1 Jamba Mini 1.6
  221. 61.0 Jamba Mini 1.5
  222. 59.8 Mixtral 8x7B
  223. 59.2 Gemini 1.5 Flash 8B
  224. 59.1 Llama 3.1-8B
  225. 59.1 Llama 3.2 11B
  226. 58.5 OLMo 2 32B Instruct
  227. 57.8 Qwen3-1.7B
  228. 57.8 Gemma 3n E2B
  229. 57.1 LFM-3B
  230. 57.1 LFM2-2.6B
  231. 57.0 Claude 3 Haiku
  232. 57.0 Mistral NeMo
  233. 56.4 Ministral 8B
  234. 56.1 Phi-3.5-mini
  235. 55.7 SmolLM3-3B
  236. 53.3 Jamba Mini 1.7
  237. 53.2 Phi-4-multimodal
  238. 52.4 Ministral 3B
  239. 52.4 Gemma 3 4B
  240. 52.3 GPT-3.5 Turbo
  241. 52.3 LFM2-8B-A1B
  242. 51.1 Phi-3.5-vision-instruct
  243. 50.4 Olmo-3 7B
  244. 49.8 Olmo-2 13B
  245. 46.0 Command R7B
  246. 42.8 LFM2.5-1.2B-Thinking
  247. 42.7 Qwen3-0.6B
  248. 41.4 LFM2-1.2B
  249. 38.5 LFM2-700M
  250. 36.5 Olmo-2 7B
  251. 33.6 Gemma 3 1B
  252. 31.8 LFM2-350M
  253. 28.0 Llama 3.2 1B
  254. 27.6 Llama 3.2 3B
  255. 27.6 SmolLM2-1.7B
  256. 27.3 LFM2.5-1.2B
  257. 25.9 SmolLM2-360M
  258. 24.3 OLMo 2 1B Instruct
  259. 21.1 SmolLM2-135M
  1. 1 Gemini 4 Argon high 92.9%
  2. 2 GPT-6 Astra high 92.7%
  3. 3 GPT-6.1 Terra medium 92.5%
  4. 4 Claude Fable 5.1 92.2%
  5. 5 GPT-5.5 92.1%
  6. 6 Gemini 3.8 Flash 91.9%
  7. 7 Claude Opus 5.5 medium 91.9%
  8. 8 GPT-6.1 Sol medium 91.8%
  9. 9 Gemini 3.1 Pro 91.5%
  10. 10 Grok 4.6 91.5%
  11. 11 GPT-5.6 Sol 91.2%
  12. 12 Claude Opus 5 91.1%
  13. 13 Gemini 3.6 Flash 90.9%
  14. 14 Gemini 3.7 Flash 90.8%
  15. 15 Gemini 3 Pro 90.5%
  16. 16 Claude Fable 5 90.5%
  17. 17 Qwen3.8 Max (0902) 90.3%
  18. 18 Qwen3.7-Max 90.1%
  19. 19 Muse Spark 1.1 90.1%
  20. 20 Claude Sonnet 4.6 Adaptive Thinking 89.7%
  21. 21 Qwen 3.6 Max (Preview) 89.7%
  22. 22 Grok 4.5 89.7%
  23. 23 Muse Wonder low 89.7%
  24. 24 Gemini 3 Flash 89.4%
  25. 25 Kimi K2.6 89.4%
  26. 26 Muse Spark 1.2 89.4%
  27. 27 Grok 4.7 low 89.4%
  28. 28 Claude Sonnet 4.5 thinking 89.3%
  29. 29 Tencent Hy4 preview 89.3%
  30. 30 Claude Opus 4.8 89.2%
  31. 31 Qwen 3.8 Max 89.2%
  32. 32 DeepSeek-V4-Pro 89.1%
  33. 33 Claude Opus 4.7 89%
  34. 34 Muse Spark 1.3 89%
  35. 35 Qwen3.5-Plus 88.9%
  36. 36 Claude Opus 4.7 Adaptive Thinking 88.8%
  37. 37 Gemini 2.5 Pro Jun 2025 88.8%
  38. 38 Qwen3.8 Flash 88.8%
  39. 39 GPT-5.4 medium 88.7%
  40. 40 Grok 4.3 88.7%
  41. 41 Grok 4 88.4%
  42. 42 Gemini 3.5 Flash 88.4%
  43. 43 Kimi K3.1 Pro 88.4%
  44. 44 Claude Sonnet 5.5 high 88.3%
  45. 45 Claude Sonnet 5 88.2%
  46. 46 Qwen3.5 397B-A17B 88.1%
  47. 47 Qwen 3.6 Plus 88.1%
  48. 48 GLM-5.1 88%
  49. 49 Qwen3.8-2.4T-A95B 88%
  50. 50 o3 87.9%
  51. 51 Claude Opus 4.6 87.9%
  52. 52 GPT-5 medium 87.8%
  53. 53 GPT-5.1 medium 87.8%
  54. 54 Kimi K2.5 87.8%
  55. 55 Kimi K3 87.8%
  56. 56 DeepSeek V4.1 Flash 87.8%
  57. 57 GLM-5 87.7%
  58. 58 Qwen3.5-27B 87.7%
  59. 59 Inkling 87.7%
  60. 60 GPT-5.2 medium 87.5%
  61. 61 Grok 4.20 reasoning 87.5%
  62. 62 Qwen3.6 27B 87.5%
  63. 63 Claude Opus 4 Thinking (low) 87.4%
  64. 64 Claude Sonnet 4.6 87.4%
  65. 65 Qwen3.5-122B-A10B 87.3%
  66. 66 GPT-5.6 Terra 87.2%
  67. 67 DeepSeek V4.1 Flash 1017 87.2%
  68. 68 Gemma 4 31B IT 87%
  69. 69 DeepSeek-V4-Flash 87%
  70. 70 GLM-5.3 87%
  71. 71 Claude Opus 4.5 86.9%
  72. 72 Claude Sonnet 4 Thinking (low) 86.8%
  73. 73 Qwen 3.8 27B 86.8%
  74. 74 Deepseek R1 (May '25) 86.7%
  75. 75 Qwen3.5-Flash 86.5%
  76. 76 o1 medium 86.4%
  77. 77 Claude Opus 4 86.3%
  78. 78 GLM-4.5 86.3%
  79. 79 Grok 4 Fast 86.3%
  80. 80 Gemini 2.5 Flash Sep 2025 86.3%
  81. 81 Nemotron 3 Ultra 86.2%
  82. 82 GLM-5.2 85.9%
  83. 83 Claude Haiku 5.5 high 85.9%
  84. 84 Qwen3-235B-A22B-Thinking Jul 2025 85.7%
  85. 85 Claude-3.7 Sonnet Thinking (medium) 85.4%
  86. 86 Doubao-Seed-1.6 85.3%
  87. 87 Gemini 3.1 Flash-Lite 85.3%
  88. 88 MiMo-V2.6 Pro 85.3%
  89. 89 GPT-4.5 85.2%
  90. 90 Mistral Large 4 85.2%
  91. 91 Grok 4.1 Fast 85.1%
  92. 92 Qwen 3.6 35B-A3B 85.1%
  93. 93 Inkling-Small 85.1%
  94. 94 Kimi K2 Thinking 84.9%
  95. 95 Claude Sonnet 4 84.8%
  96. 96 Qwen3.7 Flash 84.8%
  97. 97 GLM-4.7 84.7%
  98. 98 GPT-5.4 Mini medium 84.7%
  99. 99 MiniMax-M3 84.7%
  100. 100 Qwen3.5-35B-A3B 84.5%
  101. 101 MiMo-V2-Pro 84.1%
  102. 102 MiMo-V2.5-Pro 83.9%
  103. 103 Qwen 3.6 Flash 83.9%
  104. 104 Muse Glimmer 83.9%
  105. 105 GPT-5 mini medium 83.3%
  106. 106 Gemini 3.5 Flash-Lite 83.2%
  107. 107 MiMo-V2.6 Flash 83%
  108. 108 GLM-4.6 82.9%
  109. 109 Gemma 4 26B A4B 82.9%
  110. 110 MiMo-V2.5 82.9%
  111. 111 MiniMax-M2.7 82.8%
  112. 112 Qwen3-235B-A22B 82.4%
  113. 113 Qwen3-32B 82.4%
  114. 114 GPT-4.1 82.3%
  115. 115 GPT-5.6 Luna 82.3%
  116. 116 Mistral Medium 3.1 82.2%
  117. 117 Kimi K2 Sep 2025 82.2%
  118. 118 o4-mini 82%
  119. 119 Grok 3 81.8%
  120. 120 Mistral Medium 3.5 81.8%
  121. 121 Llama 4 Maverick 81.7%
  122. 122 Gemma 4 12B 81.7%
  123. 123 Qwen3.5-9B 81.4%
  124. 124 Mistral Large 3 81.2%
  125. 125 MiniMax-M3.1-Flash 81%
  126. 126 Doubao-1.5-pro 80.8%
  127. 127 Grok-3 mini 80.7%
  128. 128 Grok-2 Dec 2024 80.2%
  129. 129 Mistral Medium 3 80%
  130. 130 Gemini 2.0 Flash Feb 2025 79.9%
  131. 131 Qwen3-30B-A3B 79.8%
  132. 132 Claude Haiku 4.5 79.5%
  133. 133 DeepSeek-V3.2 79.5%
  134. 134 o3-mini 79.4%
  135. 135 MiniMax-M2.1 79.1%
  136. 136 Llama 3.1-405B 78.8%
  137. 137 Nemotron 3.5 Lightning 30B-A3B 78.8%
  138. 138 Qwen3-14B 78.5%
  139. 139 Seed 1.6 Flash 78.4%
  140. 140 DeepSeek-V3 Mar 2025 78.3%
  141. 141 GPT-4o Mar 2025 78.3%
  142. 142 DeepSeek-V3.1 78%
  143. 143 Command A 77.9%
  144. 144 Gemini 2.0 Flash-Lite 77.6%
  145. 145 Gemini 2.5 Flash-Lite Sep 2025 77.6%
  146. 146 LearnLM 1.5 Pro 77.3%
  147. 147 Claude 3 Opus 77%
  148. 148 Qwen2.5-32B 76.9%
  149. 149 Claude-3.5 Sonnet October 76.8%
  150. 150 Claude 3.7 Sonnet 76.8%
  151. 151 Llama 3.2 90B 76.3%
  152. 152 Olmo 3.1 32B Think 76%
  153. 153 Mistral Small 3.1 24B 75.9%
  154. 154 Command A+ 75.9%
  155. 155 Qwen2.5-14B 75.6%
  156. 156 GPT-5.4 Nano medium 75.6%
  157. 157 Qwen2.5-72B 75.4%
  158. 158 Mistral Small 3.2 75.4%
  159. 159 Mistral Large November 75.3%
  160. 160 Nova Pro 75.3%
  161. 161 Llama 3.3 70B 75.2%
  162. 162 gpt-oss-120b 75.2%
  163. 163 GPT-4.1 mini 75.1%
  164. 164 Magistral Medium 1.0 75.1%
  165. 165 Llama 3.1-70B 75%
  166. 166 GLM-4.5-Air 74.9%
  167. 167 Llama 4 Scout 74.5%
  168. 168 Gemma 4 E4B 74.4%
  169. 169 GPT-4 Turbo Apr 2024 74.3%
  170. 170 GPT-4 Jun 2023 74.1%
  171. 171 glm-4.7-flash 74%
  172. 172 Gemini 1.5 Pro Sep 2024 73.9%
  173. 173 gpt-oss-20b 73.9%
  174. 174 Nemotron 3 Nano 30B-A3B 73.8%
  175. 175 Qwen3-8B 73.5%
  176. 176 Claude-3.5 Sonnet June 73.1%
  177. 177 Grok Beta 72.9%
  178. 178 o1-mini 72.6%
  179. 179 Gemma 3 27B 72.6%
  180. 180 Mistral Small 4 72.6%
  181. 181 GPT-5 nano medium 72.4%
  182. 182 Gemini 1.5 Flash Sep 2024 72.2%
  183. 183 Phi-4 72.2%
  184. 184 Doubao-1.5 Lite 72.1%
  185. 185 LFM2.5-24B-A2B 72%
  186. 186 Yi-Lightning 71.8%
  187. 187 Yi-Large 71.5%
  188. 188 Ministral 3 14B 71.5%
  189. 189 Jamba Large 1.7 71.4%
  190. 190 Gemma 2 27B 71.2%
  191. 191 Granite 4.2 8B 71.1%
  192. 192 sonar 70.8%
  193. 193 Mistral Small 3 70.4%
  194. 194 Qwen3-4B 70.4%
  195. 195 Nova Lite 70%
  196. 196 Jamba Large 1.6 69.9%
  197. 197 Nova 2 Lite 69.6%
  198. 198 Ministral 3 8B 69.4%
  199. 199 GPT-4o mini 69.2%
  200. 200 Jamba Large 1.5 69.1%
  201. 201 Hunyuan Large 68.2%
  202. 202 Gemma 3 12B 67.6%
  203. 203 MiMo-V2-Flash 67.6%
  204. 204 Magistral Small 67.5%
  205. 205 Qwen2.5-7B 66.7%
  206. 206 Olmo-3 7B Think 66.4%
  207. 207 Mixtral 8x22B 65.8%
  208. 208 Hunyuan Large (Long Context) 65.4%
  209. 209 Olmo-3.1 32B 65.4%
  210. 210 Gemma 3n E4B 64%
  211. 211 GPT-4.1 nano 63.5%
  212. 212 Claude 3.5 Haiku 63.3%
  213. 213 Nova Micro 63.2%
  214. 214 LFM3-1.7B 62.7%
  215. 215 Gemma 4 E2B 62.6%
  216. 216 LFM-7B 61.9%
  217. 217 Command-R+ August 61.5%
  218. 218 Ministral 3 3B 61.3%
  219. 219 Jamba Instruct 61.2%
  220. 220 Jamba Mini 1.6 61.1%
  221. 221 Jamba Mini 1.5 61%
  222. 222 Mixtral 8x7B 59.8%
  223. 223 Gemini 1.5 Flash 8B 59.2%
  224. 224 Llama 3.1-8B 59.1%
  225. 225 Llama 3.2 11B 59.1%
  226. 226 OLMo 2 32B Instruct 58.5%
  227. 227 Qwen3-1.7B 57.8%
  228. 228 Gemma 3n E2B 57.8%
  229. 229 LFM-3B 57.1%
  230. 230 LFM2-2.6B 57.1%
  231. 231 Claude 3 Haiku 57%
  232. 232 Mistral NeMo 57%
  233. 233 Ministral 8B 56.4%
  234. 234 Phi-3.5-mini 56.1%
  235. 235 SmolLM3-3B 55.7%
  236. 236 Jamba Mini 1.7 53.3%
  237. 237 Phi-4-multimodal 53.2%
  238. 238 Ministral 3B 52.4%
  239. 239 Gemma 3 4B 52.4%
  240. 240 GPT-3.5 Turbo Jan 2024 52.3%
  241. 241 LFM2-8B-A1B 52.3%
  242. 242 Phi-3.5-vision-instruct 51.1%
  243. 243 Olmo-3 7B 50.4%
  244. 244 Olmo-2 13B 49.8%
  245. 245 Command R7B 46%
  246. 246 LFM2.5-1.2B-Thinking 42.8%
  247. 247 Qwen3-0.6B 42.7%
  248. 248 LFM2-1.2B 41.4%
  249. 249 LFM2-700M 38.5%
  250. 250 Olmo-2 7B 36.5%
  251. 251 Gemma 3 1B 33.6%
  252. 252 LFM2-350M 31.8%
  253. 253 Llama 3.2 1B 28%
  254. 254 Llama 3.2 3B 27.6%
  255. 255 SmolLM2-1.7B 27.6%
  256. 256 LFM2.5-1.2B 27.3%
  257. 257 SmolLM2-360M 25.9%
  258. 258 OLMo 2 1B Instruct 24.3%
  259. 259 SmolLM2-135M 21.1%

Source: Official Pedagogy Benchmark leaderboard

216 models. Scroll sideways for the rest.

  1. 92.9 Gemini 4 Argon
  2. 92.7 GPT-6 Astra
  3. 92.5 GPT-6.1 Terra
  4. 92.2 Claude Fable 5.1
  5. 92.1 GPT-5.5
  6. 91.9 Gemini 3.8 Flash
  7. 91.9 Claude Opus 5.5
  8. 91.8 GPT-6.1 Sol
  9. 91.5 Gemini 3.1 Pro
  10. 91.5 Grok 4.6
  11. 91.2 GPT-5.6 Sol
  12. 91.1 Claude Opus 5
  13. 90.9 Gemini 3.6 Flash
  14. 90.8 Gemini 3.7 Flash
  15. 90.5 Gemini 3 Pro
  16. 90.5 Claude Fable 5
  17. 90.3 Qwen3.8 Max (0902)
  18. 90.1 Qwen3.7-Max
  19. 90.1 Muse Spark 1.1
  20. 89.7 Qwen 3.6 Max (Preview)
  21. 89.7 Grok 4.5
  22. 89.7 Muse Wonder
  23. 89.4 Gemini 3 Flash
  24. 89.4 Kimi K2.6
  25. 89.4 Muse Spark 1.2
  26. 89.4 Grok 4.7
  27. 89.3 Claude Sonnet 4.5
  28. 89.3 Tencent Hy4 preview
  29. 89.2 Claude Opus 4.8
  30. 89.2 Qwen 3.8 Max
  31. 89.1 DeepSeek-V4-Pro
  32. 89.0 Claude Opus 4.7
  33. 89.0 Muse Spark 1.3
  34. 88.9 Qwen3.5-Plus
  35. 88.8 Gemini 2.5 Pro
  36. 88.8 Qwen3.8 Flash
  37. 88.7 GPT-5.4
  38. 88.7 Grok 4.3
  39. 88.4 Grok 4
  40. 88.4 Gemini 3.5 Flash
  41. 88.4 Kimi K3.1 Pro
  42. 88.3 Claude Sonnet 5.5
  43. 88.2 Claude Sonnet 5
  44. 88.1 Qwen3.5 397B-A17B
  45. 88.1 Qwen 3.6 Plus
  46. 88.0 GLM-5.1
  47. 88.0 Qwen3.8-2.4T-A95B
  48. 87.9 o3
  49. 87.9 Claude Opus 4.6
  50. 87.8 GPT-5
  51. 87.8 GPT-5.1
  52. 87.8 Kimi K2.5
  53. 87.8 Kimi K3
  54. 87.8 DeepSeek V4.1 Flash
  55. 87.7 GLM-5
  56. 87.7 Qwen3.5-27B
  57. 87.7 Inkling
  58. 87.5 GPT-5.2
  59. 87.5 Grok 4.20
  60. 87.5 Qwen3.6 27B
  61. 87.4 Claude Sonnet 4.6
  62. 87.3 Qwen3.5-122B-A10B
  63. 87.2 GPT-5.6 Terra
  64. 87.2 DeepSeek V4.1 Flash 1017
  65. 87.0 Gemma 4 31B IT
  66. 87.0 DeepSeek-V4-Flash
  67. 87.0 GLM-5.3
  68. 86.9 Claude Opus 4.5
  69. 86.8 Qwen 3.8 27B
  70. 86.5 Qwen3.5-Flash
  71. 86.4 o1
  72. 86.3 Claude Opus 4
  73. 86.3 GLM-4.5
  74. 86.3 Grok 4 Fast
  75. 86.3 Gemini 2.5 Flash
  76. 86.2 Nemotron 3 Ultra
  77. 85.9 GLM-5.2
  78. 85.9 Claude Haiku 5.5
  79. 85.7 Qwen3-235B-A22B-Thinking
  80. 85.3 Doubao-Seed-1.6
  81. 85.3 Gemini 3.1 Flash-Lite
  82. 85.3 MiMo-V2.6 Pro
  83. 85.2 GPT-4.5
  84. 85.2 Mistral Large 4
  85. 85.1 Grok 4.1 Fast
  86. 85.1 Qwen 3.6 35B-A3B
  87. 85.1 Inkling-Small
  88. 84.9 Kimi K2 Thinking
  89. 84.8 Claude Sonnet 4
  90. 84.8 Qwen3.7 Flash
  91. 84.7 GLM-4.7
  92. 84.7 GPT-5.4 Mini
  93. 84.7 MiniMax-M3
  94. 84.5 Qwen3.5-35B-A3B
  95. 84.1 MiMo-V2-Pro
  96. 83.9 MiMo-V2.5-Pro
  97. 83.9 Qwen 3.6 Flash
  98. 83.9 Muse Glimmer
  99. 83.3 GPT-5 mini
  100. 83.2 Gemini 3.5 Flash-Lite
  101. 83.0 MiMo-V2.6 Flash
  102. 82.9 GLM-4.6
  103. 82.9 Gemma 4 26B A4B
  104. 82.9 MiMo-V2.5
  105. 82.8 MiniMax-M2.7
  106. 82.4 Qwen3-235B-A22B
  107. 82.4 Qwen3-32B
  108. 82.3 GPT-4.1
  109. 82.3 GPT-5.6 Luna
  110. 82.2 Mistral Medium 3.1
  111. 82.2 Kimi K2
  112. 82.0 o4-mini
  113. 81.8 Grok 3
  114. 81.8 Mistral Medium 3.5
  115. 81.7 Llama 4 Maverick
  116. 81.7 Gemma 4 12B
  117. 81.4 Qwen3.5-9B
  118. 81.2 Mistral Large 3
  119. 81.0 MiniMax-M3.1-Flash
  120. 80.8 Doubao-1.5-pro
  121. 80.7 Grok-3 mini
  122. 80.2 Grok-2
  123. 80.0 Mistral Medium 3
  124. 79.9 Gemini 2.0 Flash
  125. 79.8 Qwen3-30B-A3B
  126. 79.5 Claude Haiku 4.5
  127. 79.5 DeepSeek-V3.2
  128. 79.4 o3-mini
  129. 79.1 MiniMax-M2.1
  130. 78.8 Llama 3.1-405B
  131. 78.8 Nemotron 3.5 Lightning 30B-A3B
  132. 78.5 Qwen3-14B
  133. 78.3 DeepSeek-V3
  134. 78.3 GPT-4o
  135. 78.0 DeepSeek-V3.1
  136. 77.6 Gemini 2.0 Flash-Lite
  137. 77.6 Gemini 2.5 Flash-Lite
  138. 77.0 Claude 3 Opus
  139. 76.9 Qwen2.5-32B
  140. 76.8 Claude 3.7 Sonnet
  141. 76.3 Llama 3.2 90B
  142. 76.0 Olmo 3.1 32B Think
  143. 75.9 Command A+
  144. 75.6 Qwen2.5-14B
  145. 75.6 GPT-5.4 Nano
  146. 75.4 Qwen2.5-72B
  147. 75.4 Mistral Small 3.2
  148. 75.2 Llama 3.3 70B
  149. 75.2 gpt-oss-120b
  150. 75.1 GPT-4.1 mini
  151. 75.1 Magistral Medium 1.0
  152. 75.0 Llama 3.1-70B
  153. 74.9 GLM-4.5-Air
  154. 74.5 Llama 4 Scout
  155. 74.4 Gemma 4 E4B
  156. 74.3 GPT-4 Turbo
  157. 74.1 GPT-4
  158. 74.0 glm-4.7-flash
  159. 73.9 Gemini 1.5 Pro
  160. 73.9 gpt-oss-20b
  161. 73.8 Nemotron 3 Nano 30B-A3B
  162. 73.5 Qwen3-8B
  163. 72.6 o1-mini
  164. 72.6 Gemma 3 27B
  165. 72.6 Mistral Small 4
  166. 72.4 GPT-5 nano
  167. 72.2 Gemini 1.5 Flash
  168. 72.2 Phi-4
  169. 72.0 LFM2.5-24B-A2B
  170. 71.8 Yi-Lightning
  171. 71.5 Yi-Large
  172. 71.5 Ministral 3 14B
  173. 71.2 Gemma 2 27B
  174. 71.1 Granite 4.2 8B
  175. 70.8 sonar
  176. 70.4 Mistral Small 3
  177. 70.4 Qwen3-4B
  178. 69.6 Nova 2 Lite
  179. 69.2 GPT-4o mini
  180. 67.6 Gemma 3 12B
  181. 67.6 MiMo-V2-Flash
  182. 66.7 Qwen2.5-7B
  183. 65.8 Mixtral 8x22B
  184. 64.0 Gemma 3n E4B
  185. 63.5 GPT-4.1 nano
  186. 63.3 Claude 3.5 Haiku
  187. 62.7 LFM3-1.7B
  188. 62.6 Gemma 4 E2B
  189. 59.8 Mixtral 8x7B
  190. 59.2 Gemini 1.5 Flash 8B
  191. 59.1 Llama 3.1-8B
  192. 59.1 Llama 3.2 11B
  193. 58.5 OLMo 2 32B Instruct
  194. 57.8 Qwen3-1.7B
  195. 57.8 Gemma 3n E2B
  196. 57.0 Claude 3 Haiku
  197. 57.0 Mistral NeMo
  198. 56.4 Ministral 8B
  199. 56.1 Phi-3.5-mini
  200. 53.2 Phi-4-multimodal
  201. 52.4 Ministral 3B
  202. 52.4 Gemma 3 4B
  203. 52.3 GPT-3.5 Turbo
  204. 52.3 LFM2-8B-A1B
  205. 51.1 Phi-3.5-vision-instruct
  206. 46.0 Command R7B
  207. 42.8 LFM2.5-1.2B-Thinking
  208. 42.7 Qwen3-0.6B
  209. 41.4 LFM2-1.2B
  210. 33.6 Gemma 3 1B
  211. 28.0 Llama 3.2 1B
  212. 27.6 Llama 3.2 3B
  213. 27.6 SmolLM2-1.7B
  214. 25.9 SmolLM2-360M
  215. 24.3 OLMo 2 1B Instruct
  216. 21.1 SmolLM2-135M
  1. 1 Gemini 4 Argon high 92.9%
  2. 2 GPT-6 Astra high 92.7%
  3. 3 GPT-6.1 Terra medium 92.5%
  4. 4 Claude Fable 5.1 92.2%
  5. 5 GPT-5.5 92.1%
  6. 6 Gemini 3.8 Flash 91.9%
  7. 7 Claude Opus 5.5 medium 91.9%
  8. 8 GPT-6.1 Sol medium 91.8%
  9. 9 Gemini 3.1 Pro 91.5%
  10. 10 Grok 4.6 91.5%
  11. 11 GPT-5.6 Sol 91.2%
  12. 12 Claude Opus 5 91.1%
  13. 13 Gemini 3.6 Flash 90.9%
  14. 14 Gemini 3.7 Flash 90.8%
  15. 15 Gemini 3 Pro 90.5%
  16. 16 Claude Fable 5 90.5%
  17. 17 Qwen3.8 Max (0902) 90.3%
  18. 18 Qwen3.7-Max 90.1%
  19. 19 Muse Spark 1.1 90.1%
  20. 20 Qwen 3.6 Max (Preview) 89.7%
  21. 21 Grok 4.5 89.7%
  22. 22 Muse Wonder low 89.7%
  23. 23 Gemini 3 Flash 89.4%
  24. 24 Kimi K2.6 89.4%
  25. 25 Muse Spark 1.2 89.4%
  26. 26 Grok 4.7 low 89.4%
  27. 27 Claude Sonnet 4.5 thinking 89.3%
  28. 28 Tencent Hy4 preview 89.3%
  29. 29 Claude Opus 4.8 89.2%
  30. 30 Qwen 3.8 Max 89.2%
  31. 31 DeepSeek-V4-Pro 89.1%
  32. 32 Claude Opus 4.7 89%
  33. 33 Muse Spark 1.3 89%
  34. 34 Qwen3.5-Plus 88.9%
  35. 35 Gemini 2.5 Pro Jun 2025 88.8%
  36. 36 Qwen3.8 Flash 88.8%
  37. 37 GPT-5.4 medium 88.7%
  38. 38 Grok 4.3 88.7%
  39. 39 Grok 4 88.4%
  40. 40 Gemini 3.5 Flash 88.4%
  41. 41 Kimi K3.1 Pro 88.4%
  42. 42 Claude Sonnet 5.5 high 88.3%
  43. 43 Claude Sonnet 5 88.2%
  44. 44 Qwen3.5 397B-A17B 88.1%
  45. 45 Qwen 3.6 Plus 88.1%
  46. 46 GLM-5.1 88%
  47. 47 Qwen3.8-2.4T-A95B 88%
  48. 48 o3 87.9%
  49. 49 Claude Opus 4.6 87.9%
  50. 50 GPT-5 medium 87.8%
  51. 51 GPT-5.1 medium 87.8%
  52. 52 Kimi K2.5 87.8%
  53. 53 Kimi K3 87.8%
  54. 54 DeepSeek V4.1 Flash 87.8%
  55. 55 GLM-5 87.7%
  56. 56 Qwen3.5-27B 87.7%
  57. 57 Inkling 87.7%
  58. 58 GPT-5.2 medium 87.5%
  59. 59 Grok 4.20 reasoning 87.5%
  60. 60 Qwen3.6 27B 87.5%
  61. 61 Claude Sonnet 4.6 87.4%
  62. 62 Qwen3.5-122B-A10B 87.3%
  63. 63 GPT-5.6 Terra 87.2%
  64. 64 DeepSeek V4.1 Flash 1017 87.2%
  65. 65 Gemma 4 31B IT 87%
  66. 66 DeepSeek-V4-Flash 87%
  67. 67 GLM-5.3 87%
  68. 68 Claude Opus 4.5 86.9%
  69. 69 Qwen 3.8 27B 86.8%
  70. 70 Qwen3.5-Flash 86.5%
  71. 71 o1 medium 86.4%
  72. 72 Claude Opus 4 86.3%
  73. 73 GLM-4.5 86.3%
  74. 74 Grok 4 Fast 86.3%
  75. 75 Gemini 2.5 Flash Sep 2025 86.3%
  76. 76 Nemotron 3 Ultra 86.2%
  77. 77 GLM-5.2 85.9%
  78. 78 Claude Haiku 5.5 high 85.9%
  79. 79 Qwen3-235B-A22B-Thinking Jul 2025 85.7%
  80. 80 Doubao-Seed-1.6 85.3%
  81. 81 Gemini 3.1 Flash-Lite 85.3%
  82. 82 MiMo-V2.6 Pro 85.3%
  83. 83 GPT-4.5 85.2%
  84. 84 Mistral Large 4 85.2%
  85. 85 Grok 4.1 Fast 85.1%
  86. 86 Qwen 3.6 35B-A3B 85.1%
  87. 87 Inkling-Small 85.1%
  88. 88 Kimi K2 Thinking 84.9%
  89. 89 Claude Sonnet 4 84.8%
  90. 90 Qwen3.7 Flash 84.8%
  91. 91 GLM-4.7 84.7%
  92. 92 GPT-5.4 Mini medium 84.7%
  93. 93 MiniMax-M3 84.7%
  94. 94 Qwen3.5-35B-A3B 84.5%
  95. 95 MiMo-V2-Pro 84.1%
  96. 96 MiMo-V2.5-Pro 83.9%
  97. 97 Qwen 3.6 Flash 83.9%
  98. 98 Muse Glimmer 83.9%
  99. 99 GPT-5 mini medium 83.3%
  100. 100 Gemini 3.5 Flash-Lite 83.2%
  101. 101 MiMo-V2.6 Flash 83%
  102. 102 GLM-4.6 82.9%
  103. 103 Gemma 4 26B A4B 82.9%
  104. 104 MiMo-V2.5 82.9%
  105. 105 MiniMax-M2.7 82.8%
  106. 106 Qwen3-235B-A22B 82.4%
  107. 107 Qwen3-32B 82.4%
  108. 108 GPT-4.1 82.3%
  109. 109 GPT-5.6 Luna 82.3%
  110. 110 Mistral Medium 3.1 82.2%
  111. 111 Kimi K2 Sep 2025 82.2%
  112. 112 o4-mini 82%
  113. 113 Grok 3 81.8%
  114. 114 Mistral Medium 3.5 81.8%
  115. 115 Llama 4 Maverick 81.7%
  116. 116 Gemma 4 12B 81.7%
  117. 117 Qwen3.5-9B 81.4%
  118. 118 Mistral Large 3 81.2%
  119. 119 MiniMax-M3.1-Flash 81%
  120. 120 Doubao-1.5-pro 80.8%
  121. 121 Grok-3 mini 80.7%
  122. 122 Grok-2 Dec 2024 80.2%
  123. 123 Mistral Medium 3 80%
  124. 124 Gemini 2.0 Flash Feb 2025 79.9%
  125. 125 Qwen3-30B-A3B 79.8%
  126. 126 Claude Haiku 4.5 79.5%
  127. 127 DeepSeek-V3.2 79.5%
  128. 128 o3-mini 79.4%
  129. 129 MiniMax-M2.1 79.1%
  130. 130 Llama 3.1-405B 78.8%
  131. 131 Nemotron 3.5 Lightning 30B-A3B 78.8%
  132. 132 Qwen3-14B 78.5%
  133. 133 DeepSeek-V3 Mar 2025 78.3%
  134. 134 GPT-4o Mar 2025 78.3%
  135. 135 DeepSeek-V3.1 78%
  136. 136 Gemini 2.0 Flash-Lite 77.6%
  137. 137 Gemini 2.5 Flash-Lite Sep 2025 77.6%
  138. 138 Claude 3 Opus 77%
  139. 139 Qwen2.5-32B 76.9%
  140. 140 Claude 3.7 Sonnet 76.8%
  141. 141 Llama 3.2 90B 76.3%
  142. 142 Olmo 3.1 32B Think 76%
  143. 143 Command A+ 75.9%
  144. 144 Qwen2.5-14B 75.6%
  145. 145 GPT-5.4 Nano medium 75.6%
  146. 146 Qwen2.5-72B 75.4%
  147. 147 Mistral Small 3.2 75.4%
  148. 148 Llama 3.3 70B 75.2%
  149. 149 gpt-oss-120b 75.2%
  150. 150 GPT-4.1 mini 75.1%
  151. 151 Magistral Medium 1.0 75.1%
  152. 152 Llama 3.1-70B 75%
  153. 153 GLM-4.5-Air 74.9%
  154. 154 Llama 4 Scout 74.5%
  155. 155 Gemma 4 E4B 74.4%
  156. 156 GPT-4 Turbo Apr 2024 74.3%
  157. 157 GPT-4 Jun 2023 74.1%
  158. 158 glm-4.7-flash 74%
  159. 159 Gemini 1.5 Pro Sep 2024 73.9%
  160. 160 gpt-oss-20b 73.9%
  161. 161 Nemotron 3 Nano 30B-A3B 73.8%
  162. 162 Qwen3-8B 73.5%
  163. 163 o1-mini 72.6%
  164. 164 Gemma 3 27B 72.6%
  165. 165 Mistral Small 4 72.6%
  166. 166 GPT-5 nano medium 72.4%
  167. 167 Gemini 1.5 Flash Sep 2024 72.2%
  168. 168 Phi-4 72.2%
  169. 169 LFM2.5-24B-A2B 72%
  170. 170 Yi-Lightning 71.8%
  171. 171 Yi-Large 71.5%
  172. 172 Ministral 3 14B 71.5%
  173. 173 Gemma 2 27B 71.2%
  174. 174 Granite 4.2 8B 71.1%
  175. 175 sonar 70.8%
  176. 176 Mistral Small 3 70.4%
  177. 177 Qwen3-4B 70.4%
  178. 178 Nova 2 Lite 69.6%
  179. 179 GPT-4o mini 69.2%
  180. 180 Gemma 3 12B 67.6%
  181. 181 MiMo-V2-Flash 67.6%
  182. 182 Qwen2.5-7B 66.7%
  183. 183 Mixtral 8x22B 65.8%
  184. 184 Gemma 3n E4B 64%
  185. 185 GPT-4.1 nano 63.5%
  186. 186 Claude 3.5 Haiku 63.3%
  187. 187 LFM3-1.7B 62.7%
  188. 188 Gemma 4 E2B 62.6%
  189. 189 Mixtral 8x7B 59.8%
  190. 190 Gemini 1.5 Flash 8B 59.2%
  191. 191 Llama 3.1-8B 59.1%
  192. 192 Llama 3.2 11B 59.1%
  193. 193 OLMo 2 32B Instruct 58.5%
  194. 194 Qwen3-1.7B 57.8%
  195. 195 Gemma 3n E2B 57.8%
  196. 196 Claude 3 Haiku 57%
  197. 197 Mistral NeMo 57%
  198. 198 Ministral 8B 56.4%
  199. 199 Phi-3.5-mini 56.1%
  200. 200 Phi-4-multimodal 53.2%
  201. 201 Ministral 3B 52.4%
  202. 202 Gemma 3 4B 52.4%
  203. 203 GPT-3.5 Turbo Jan 2024 52.3%
  204. 204 LFM2-8B-A1B 52.3%
  205. 205 Phi-3.5-vision-instruct 51.1%
  206. 206 Command R7B 46%
  207. 207 LFM2.5-1.2B-Thinking 42.8%
  208. 208 Qwen3-0.6B 42.7%
  209. 209 LFM2-1.2B 41.4%
  210. 210 Gemma 3 1B 33.6%
  211. 211 Llama 3.2 1B 28%
  212. 212 Llama 3.2 3B 27.6%
  213. 213 SmolLM2-1.7B 27.6%
  214. 214 SmolLM2-360M 25.9%
  215. 215 OLMo 2 1B Instruct 24.3%
  216. 216 SmolLM2-135M 21.1%

Source: Official Pedagogy Benchmark leaderboard

#EntryScoreSettingDateSource
1 Gemini 4 Argon 92.9% high Sep 2026 Official leaderboard
2 GPT-6 Astra 92.7% high Sep 2026 Official leaderboard
3 GPT-6.1 Terra 92.5% medium Oct 2026 Official leaderboard
4 Claude Fable 5.1 92.2% Sep 2026 Official leaderboard
5 GPT-5.5 92.1% Apr 2026 Official leaderboard
6 Gemini 3.8 Flash 91.9% Sep 2026 Official leaderboard
7 Claude Opus 5.5 91.9% medium Sep 2026 Official leaderboard
8 GPT-6.1 Sol 91.8% medium Sep 2026 Official leaderboard
9 Gemini 3.1 Pro 91.5% Feb 2026 Official leaderboard
10 Grok 4.6 91.5% Aug 2026 Official leaderboard
11 GPT-5.6 Sol 91.2% Jul 2026 Official leaderboard
12 Claude Opus 5 91.1% Jul 2026 Official leaderboard
13 Gemini 3.6 Flash 90.9% Jul 2026 Official leaderboard
14 Gemini 3.7 Flash 90.8% Aug 2026 Official leaderboard
15 Gemini 3 Pro 90.5% Nov 2025 Official leaderboard
16 Claude Fable 5 90.5% Jun 2026 Official leaderboard
17 Qwen3.8 Max (0902) 90.3% Sep 2026 Official leaderboard
18 Qwen3.7-Max 90.1% May 2026 Official leaderboard
19 Muse Spark 1.1 90.1% Jul 2026 Official leaderboard
20 Claude Sonnet 4.6 Adaptive Thinking 89.7% Official leaderboard
21 Qwen 3.6 Max (Preview) 89.7% Apr 2026 Official leaderboard
22 Grok 4.5 89.7% Jul 2026 Official leaderboard
23 Muse Wonder 89.7% low Oct 2026 Official leaderboard
24 Gemini 3 Flash 89.4% Dec 2025 Official leaderboard
25 Kimi K2.6 89.4% Apr 2026 Official leaderboard
26 Muse Spark 1.2 89.4% Aug 2026 Official leaderboard
27 Grok 4.7 89.4% low Sep 2026 Official leaderboard
28 Claude Sonnet 4.5 89.3% thinking Sep 2025 Official leaderboard
29 Tencent Hy4 preview 89.3% Aug 2026 Official leaderboard
30 Claude Opus 4.8 89.2% May 2026 Official leaderboard
31 Qwen 3.8 Max 89.2% Aug 2026 Official leaderboard
32 DeepSeek-V4-Pro 89.1% Apr 2026 Official leaderboard
33 Claude Opus 4.7 89% Apr 2026 Official leaderboard
34 Muse Spark 1.3 89% Sep 2026 Official leaderboard
35 Qwen3.5-Plus 88.9% Feb 2026 Official leaderboard
36 Claude Opus 4.7 Adaptive Thinking 88.8% Official leaderboard
37 Gemini 2.5 Pro Jun 2025 88.8% Jun 2025 Official leaderboard
38 Qwen3.8 Flash 88.8% Aug 2026 Official leaderboard
39 GPT-5.4 88.7% medium Mar 2026 Official leaderboard
40 Grok 4.3 88.7% Apr 2026 Official leaderboard
41 Grok 4 88.4% Jul 2025 Official leaderboard
42 Gemini 3.5 Flash 88.4% May 2026 Official leaderboard
43 Kimi K3.1 Pro 88.4% Oct 2026 Official leaderboard
44 Claude Sonnet 5.5 88.3% high Sep 2026 Official leaderboard
45 Claude Sonnet 5 88.2% Jun 2026 Official leaderboard
46 Qwen3.5 397B-A17B 88.1% Feb 2026 Official leaderboard
47 Qwen 3.6 Plus 88.1% Mar 2026 Official leaderboard
48 GLM-5.1 88% Apr 2026 Official leaderboard
49 Qwen3.8-2.4T-A95B 88% Aug 2026 Official leaderboard
50 o3 87.9% Apr 2025 Official leaderboard
51 Claude Opus 4.6 87.9% Feb 2026 Official leaderboard
52 GPT-5 87.8% medium Aug 2025 Official leaderboard
53 GPT-5.1 87.8% medium Nov 2025 Official leaderboard
54 Kimi K2.5 87.8% Jan 2026 Official leaderboard
55 Kimi K3 87.8% Jul 2026 Official leaderboard
56 DeepSeek V4.1 Flash 87.8% Sep 2026 Official leaderboard
57 GLM-5 87.7% Feb 2026 Official leaderboard
58 Qwen3.5-27B 87.7% Feb 2026 Official leaderboard
59 Inkling 87.7% Jul 2026 Official leaderboard
60 GPT-5.2 87.5% medium Dec 2025 Official leaderboard
61 Grok 4.20 87.5% reasoning Feb 2026 Official leaderboard
62 Qwen3.6 27B 87.5% Apr 2026 Official leaderboard
63 Claude Opus 4 Thinking (low) 87.4% Official leaderboard
64 Claude Sonnet 4.6 87.4% Feb 2026 Official leaderboard
65 Qwen3.5-122B-A10B 87.3% Feb 2026 Official leaderboard
66 GPT-5.6 Terra 87.2% Jul 2026 Official leaderboard
67 DeepSeek V4.1 Flash 1017 87.2% Oct 2026 Official leaderboard
68 Gemma 4 31B IT 87% Apr 2026 Official leaderboard
69 DeepSeek-V4-Flash 87% Apr 2026 Official leaderboard
70 GLM-5.3 87% Aug 2026 Official leaderboard
71 Claude Opus 4.5 86.9% Nov 2025 Official leaderboard
72 Claude Sonnet 4 Thinking (low) 86.8% Official leaderboard
73 Qwen 3.8 27B 86.8% Aug 2026 Official leaderboard
74 Deepseek R1 (May '25) 86.7% Official leaderboard
75 Qwen3.5-Flash 86.5% Feb 2026 Official leaderboard
76 o1 86.4% medium Dec 2024 Official leaderboard
77 Claude Opus 4 86.3% May 2025 Official leaderboard
78 GLM-4.5 86.3% Aug 2025 Official leaderboard
79 Grok 4 Fast 86.3% Sep 2025 Official leaderboard
80 Gemini 2.5 Flash Sep 2025 86.3% Sep 2025 Official leaderboard
81 Nemotron 3 Ultra 86.2% Jun 2026 Official leaderboard
82 GLM-5.2 85.9% Jun 2026 Official leaderboard
83 Claude Haiku 5.5 85.9% high Oct 2026 Official leaderboard
84 Qwen3-235B-A22B-Thinking Jul 2025 85.7% Jul 2025 Official leaderboard
85 Claude-3.7 Sonnet Thinking (medium) 85.4% Official leaderboard
86 Doubao-Seed-1.6 85.3% Jun 2025 Official leaderboard
87 Gemini 3.1 Flash-Lite 85.3% Mar 2026 Official leaderboard
88 MiMo-V2.6 Pro 85.3% Sep 2026 Official leaderboard
89 GPT-4.5 85.2% Feb 2025 Official leaderboard
90 Mistral Large 4 85.2% Oct 2026 Official leaderboard
91 Grok 4.1 Fast 85.1% Nov 2025 Official leaderboard
92 Qwen 3.6 35B-A3B 85.1% Apr 2026 Official leaderboard
93 Inkling-Small 85.1% Jul 2026 Official leaderboard
94 Kimi K2 Thinking 84.9% Nov 2025 Official leaderboard
95 Claude Sonnet 4 84.8% May 2025 Official leaderboard
96 Qwen3.7 Flash 84.8% Jul 2026 Official leaderboard
97 GLM-4.7 84.7% Dec 2025 Official leaderboard
98 GPT-5.4 Mini 84.7% medium Mar 2026 Official leaderboard
99 MiniMax-M3 84.7% Jun 2026 Official leaderboard
100 Qwen3.5-35B-A3B 84.5% Feb 2026 Official leaderboard
101 MiMo-V2-Pro 84.1% Mar 2026 Official leaderboard
102 MiMo-V2.5-Pro 83.9% Apr 2026 Official leaderboard
103 Qwen 3.6 Flash 83.9% Apr 2026 Official leaderboard
104 Muse Glimmer 83.9% Aug 2026 Official leaderboard
105 GPT-5 mini 83.3% medium Aug 2025 Official leaderboard
106 Gemini 3.5 Flash-Lite 83.2% Jul 2026 Official leaderboard
107 MiMo-V2.6 Flash 83% Sep 2026 Official leaderboard
108 GLM-4.6 82.9% Sep 2025 Official leaderboard
109 Gemma 4 26B A4B 82.9% Apr 2026 Official leaderboard
110 MiMo-V2.5 82.9% Apr 2026 Official leaderboard
111 MiniMax-M2.7 82.8% Mar 2026 Official leaderboard
112 Qwen3-235B-A22B 82.4% Apr 2025 Official leaderboard
113 Qwen3-32B 82.4% Apr 2025 Official leaderboard
114 GPT-4.1 82.3% Apr 2025 Official leaderboard
115 GPT-5.6 Luna 82.3% Jul 2026 Official leaderboard
116 Mistral Medium 3.1 82.2% Aug 2025 Official leaderboard
117 Kimi K2 Sep 2025 82.2% Sep 2025 Official leaderboard
118 o4-mini 82% Apr 2025 Official leaderboard
119 Grok 3 81.8% Apr 2025 Official leaderboard
120 Mistral Medium 3.5 81.8% Apr 2026 Official leaderboard
121 Llama 4 Maverick 81.7% Apr 2025 Official leaderboard
122 Gemma 4 12B 81.7% Jun 2026 Official leaderboard
123 Qwen3.5-9B 81.4% Feb 2026 Official leaderboard
124 Mistral Large 3 81.2% Dec 2025 Official leaderboard
125 MiniMax-M3.1-Flash 81% Oct 2026 Official leaderboard
126 Doubao-1.5-pro 80.8% Jan 2025 Official leaderboard
127 Grok-3 mini 80.7% Apr 2025 Official leaderboard
128 Grok-2 Dec 2024 80.2% Dec 2024 Official leaderboard
129 Mistral Medium 3 80% May 2025 Official leaderboard
130 Gemini 2.0 Flash Feb 2025 79.9% Feb 2025 Official leaderboard
131 Qwen3-30B-A3B 79.8% Apr 2025 Official leaderboard
132 Claude Haiku 4.5 79.5% Oct 2025 Official leaderboard
133 DeepSeek-V3.2 79.5% Dec 2025 Official leaderboard
134 o3-mini 79.4% Jan 2025 Official leaderboard
135 MiniMax-M2.1 79.1% Dec 2025 Official leaderboard
136 Llama 3.1-405B 78.8% Jul 2024 Official leaderboard
137 Nemotron 3.5 Lightning 30B-A3B 78.8% Aug 2026 Official leaderboard
138 Qwen3-14B 78.5% Apr 2025 Official leaderboard
139 Seed 1.6 Flash 78.4% Official leaderboard
140 DeepSeek-V3 Mar 2025 78.3% Mar 2025 Official leaderboard
141 GPT-4o Mar 2025 78.3% Mar 2025 Official leaderboard
142 DeepSeek-V3.1 78% Aug 2025 Official leaderboard
143 Command A 77.9% Official leaderboard
144 Gemini 2.0 Flash-Lite 77.6% Feb 2025 Official leaderboard
145 Gemini 2.5 Flash-Lite Sep 2025 77.6% Sep 2025 Official leaderboard
146 LearnLM 1.5 Pro 77.3% Official leaderboard
147 Claude 3 Opus 77% Feb 2024 Official leaderboard
148 Qwen2.5-32B 76.9% Sep 2024 Official leaderboard
149 Claude-3.5 Sonnet October 76.8% Official leaderboard
150 Claude 3.7 Sonnet 76.8% Feb 2025 Official leaderboard
151 Llama 3.2 90B 76.3% Sep 2024 Official leaderboard
152 Olmo 3.1 32B Think 76% Dec 2025 Official leaderboard
153 Mistral Small 3.1 24B 75.9% Official leaderboard
154 Command A+ 75.9% May 2026 Official leaderboard
155 Qwen2.5-14B 75.6% Sep 2024 Official leaderboard
156 GPT-5.4 Nano 75.6% medium Mar 2026 Official leaderboard
157 Qwen2.5-72B 75.4% Sep 2024 Official leaderboard
158 Mistral Small 3.2 75.4% Jun 2025 Official leaderboard
159 Mistral Large November 75.3% Official leaderboard
160 Nova Pro 75.3% Official leaderboard
161 Llama 3.3 70B 75.2% Dec 2024 Official leaderboard
162 gpt-oss-120b 75.2% Aug 2025 Official leaderboard
163 GPT-4.1 mini 75.1% Apr 2025 Official leaderboard
164 Magistral Medium 1.0 75.1% Jun 2025 Official leaderboard
165 Llama 3.1-70B 75% Jul 2024 Official leaderboard
166 GLM-4.5-Air 74.9% Jul 2025 Official leaderboard
167 Llama 4 Scout 74.5% Apr 2025 Official leaderboard
168 Gemma 4 E4B 74.4% Mar 2026 Official leaderboard
169 GPT-4 Turbo Apr 2024 74.3% Apr 2024 Official leaderboard
170 GPT-4 Jun 2023 74.1% Jun 2023 Official leaderboard
171 glm-4.7-flash 74% Jan 2026 Official leaderboard
172 Gemini 1.5 Pro Sep 2024 73.9% Sep 2024 Official leaderboard
173 gpt-oss-20b 73.9% Aug 2025 Official leaderboard
174 Nemotron 3 Nano 30B-A3B 73.8% Dec 2025 Official leaderboard
175 Qwen3-8B 73.5% Apr 2025 Official leaderboard
176 Claude-3.5 Sonnet June 73.1% Official leaderboard
177 Grok Beta 72.9% Official leaderboard
178 o1-mini 72.6% Sep 2024 Official leaderboard
179 Gemma 3 27B 72.6% Mar 2025 Official leaderboard
180 Mistral Small 4 72.6% Mar 2026 Official leaderboard
181 GPT-5 nano 72.4% medium Aug 2025 Official leaderboard
182 Gemini 1.5 Flash Sep 2024 72.2% Aug 2024 Official leaderboard
183 Phi-4 72.2% Dec 2024 Official leaderboard
184 Doubao-1.5 Lite 72.1% Official leaderboard
185 LFM2.5-24B-A2B 72% Oct 2026 Official leaderboard
186 Yi-Lightning 71.8% Dec 2024 Official leaderboard
187 Yi-Large 71.5% May 2024 Official leaderboard
188 Ministral 3 14B 71.5% Dec 2025 Official leaderboard
189 Jamba Large 1.7 71.4% Official leaderboard
190 Gemma 2 27B 71.2% Jun 2024 Official leaderboard
191 Granite 4.2 8B 71.1% Aug 2026 Official leaderboard
192 sonar 70.8% Jan 2025 Official leaderboard
193 Mistral Small 3 70.4% Jan 2025 Official leaderboard
194 Qwen3-4B 70.4% Apr 2025 Official leaderboard
195 Nova Lite 70% Official leaderboard
196 Jamba Large 1.6 69.9% Official leaderboard
197 Nova 2 Lite 69.6% Dec 2025 Official leaderboard
198 Ministral 3 8B 69.4% Official leaderboard
199 GPT-4o mini 69.2% Jul 2024 Official leaderboard
200 Jamba Large 1.5 69.1% Official leaderboard
201 Hunyuan Large 68.2% Official leaderboard
202 Gemma 3 12B 67.6% Mar 2025 Official leaderboard
203 MiMo-V2-Flash 67.6% Dec 2025 Official leaderboard
204 Magistral Small 67.5% Official leaderboard
205 Qwen2.5-7B 66.7% Sep 2024 Official leaderboard
206 Olmo-3 7B Think 66.4% Official leaderboard
207 Mixtral 8x22B 65.8% Apr 2024 Official leaderboard
208 Hunyuan Large (Long Context) 65.4% Official leaderboard
209 Olmo-3.1 32B 65.4% Official leaderboard
210 Gemma 3n E4B 64% Jun 2025 Official leaderboard
211 GPT-4.1 nano 63.5% Apr 2025 Official leaderboard
212 Claude 3.5 Haiku 63.3% Oct 2024 Official leaderboard
213 Nova Micro 63.2% Official leaderboard
214 LFM3-1.7B 62.7% Oct 2026 Official leaderboard
215 Gemma 4 E2B 62.6% Mar 2026 Official leaderboard
216 LFM-7B 61.9% Official leaderboard
217 Command-R+ August 61.5% Official leaderboard
218 Ministral 3 3B 61.3% Official leaderboard
219 Jamba Instruct 61.2% Official leaderboard
220 Jamba Mini 1.6 61.1% Official leaderboard
221 Jamba Mini 1.5 61% Official leaderboard
222 Mixtral 8x7B 59.8% Dec 2023 Official leaderboard
223 Gemini 1.5 Flash 8B 59.2% Oct 2024 Official leaderboard
224 Llama 3.1-8B 59.1% Jul 2024 Official leaderboard
225 Llama 3.2 11B 59.1% Sep 2024 Official leaderboard
226 OLMo 2 32B Instruct 58.5% Mar 2025 Official leaderboard
227 Qwen3-1.7B 57.8% Apr 2025 Official leaderboard
228 Gemma 3n E2B 57.8% Jun 2025 Official leaderboard
229 LFM-3B 57.1% Official leaderboard
230 LFM2-2.6B 57.1% Official leaderboard
231 Claude 3 Haiku 57% Mar 2024 Official leaderboard
232 Mistral NeMo 57% Jul 2024 Official leaderboard
233 Ministral 8B 56.4% Oct 2024 Official leaderboard
234 Phi-3.5-mini 56.1% Aug 2024 Official leaderboard
235 SmolLM3-3B 55.7% Official leaderboard
236 Jamba Mini 1.7 53.3% Official leaderboard
237 Phi-4-multimodal 53.2% Feb 2025 Official leaderboard
238 Ministral 3B 52.4% Oct 2024 Official leaderboard
239 Gemma 3 4B 52.4% Mar 2025 Official leaderboard
240 GPT-3.5 Turbo Jan 2024 52.3% Jan 2024 Official leaderboard
241 LFM2-8B-A1B 52.3% Oct 2025 Official leaderboard
242 Phi-3.5-vision-instruct 51.1% Aug 2024 Official leaderboard
243 Olmo-3 7B 50.4% Official leaderboard
244 Olmo-2 13B 49.8% Official leaderboard
245 Command R7B 46% Dec 2024 Official leaderboard
246 LFM2.5-1.2B-Thinking 42.8% Jan 2026 Official leaderboard
247 Qwen3-0.6B 42.7% Apr 2025 Official leaderboard
248 LFM2-1.2B 41.4% Jul 2025 Official leaderboard
249 LFM2-700M 38.5% Official leaderboard
250 Olmo-2 7B 36.5% Official leaderboard
251 Gemma 3 1B 33.6% Mar 2025 Official leaderboard
252 LFM2-350M 31.8% Official leaderboard
253 Llama 3.2 1B 28% Sep 2024 Official leaderboard
254 Llama 3.2 3B 27.6% Sep 2024 Official leaderboard
255 SmolLM2-1.7B 27.6% Oct 2024 Official leaderboard
256 LFM2.5-1.2B 27.3% Official leaderboard
257 SmolLM2-360M 25.9% Oct 2024 Official leaderboard
258 OLMo 2 1B Instruct 24.3% Apr 2025 Official leaderboard
259 SmolLM2-135M 21.1% Oct 2024 Official leaderboard
#ModelScoreSettingDateSource
1 Gemini 4 Argon 92.9% high Sep 2026 Official leaderboard
2 GPT-6 Astra 92.7% high Sep 2026 Official leaderboard
3 GPT-6.1 Terra 92.5% medium Oct 2026 Official leaderboard
4 Claude Fable 5.1 92.2% Sep 2026 Official leaderboard
5 GPT-5.5 92.1% Apr 2026 Official leaderboard
6 Gemini 3.8 Flash 91.9% Sep 2026 Official leaderboard
7 Claude Opus 5.5 91.9% medium Sep 2026 Official leaderboard
8 GPT-6.1 Sol 91.8% medium Sep 2026 Official leaderboard
9 Gemini 3.1 Pro 91.5% Feb 2026 Official leaderboard
10 Grok 4.6 91.5% Aug 2026 Official leaderboard
11 GPT-5.6 Sol 91.2% Jul 2026 Official leaderboard
12 Claude Opus 5 91.1% Jul 2026 Official leaderboard
13 Gemini 3.6 Flash 90.9% Jul 2026 Official leaderboard
14 Gemini 3.7 Flash 90.8% Aug 2026 Official leaderboard
15 Gemini 3 Pro 90.5% Nov 2025 Official leaderboard
16 Claude Fable 5 90.5% Jun 2026 Official leaderboard
17 Qwen3.8 Max (0902) 90.3% Sep 2026 Official leaderboard
18 Qwen3.7-Max 90.1% May 2026 Official leaderboard
19 Muse Spark 1.1 90.1% Jul 2026 Official leaderboard
20 Qwen 3.6 Max (Preview) 89.7% Apr 2026 Official leaderboard
21 Grok 4.5 89.7% Jul 2026 Official leaderboard
22 Muse Wonder 89.7% low Oct 2026 Official leaderboard
23 Gemini 3 Flash 89.4% Dec 2025 Official leaderboard
24 Kimi K2.6 89.4% Apr 2026 Official leaderboard
25 Muse Spark 1.2 89.4% Aug 2026 Official leaderboard
26 Grok 4.7 89.4% low Sep 2026 Official leaderboard
27 Claude Sonnet 4.5 89.3% thinking Sep 2025 Official leaderboard
28 Tencent Hy4 preview 89.3% Aug 2026 Official leaderboard
29 Claude Opus 4.8 89.2% May 2026 Official leaderboard
30 Qwen 3.8 Max 89.2% Aug 2026 Official leaderboard
31 DeepSeek-V4-Pro 89.1% Apr 2026 Official leaderboard
32 Claude Opus 4.7 89% Apr 2026 Official leaderboard
33 Muse Spark 1.3 89% Sep 2026 Official leaderboard
34 Qwen3.5-Plus 88.9% Feb 2026 Official leaderboard
35 Gemini 2.5 Pro Jun 2025 88.8% Jun 2025 Official leaderboard
36 Qwen3.8 Flash 88.8% Aug 2026 Official leaderboard
37 GPT-5.4 88.7% medium Mar 2026 Official leaderboard
38 Grok 4.3 88.7% Apr 2026 Official leaderboard
39 Grok 4 88.4% Jul 2025 Official leaderboard
40 Gemini 3.5 Flash 88.4% May 2026 Official leaderboard
41 Kimi K3.1 Pro 88.4% Oct 2026 Official leaderboard
42 Claude Sonnet 5.5 88.3% high Sep 2026 Official leaderboard
43 Claude Sonnet 5 88.2% Jun 2026 Official leaderboard
44 Qwen3.5 397B-A17B 88.1% Feb 2026 Official leaderboard
45 Qwen 3.6 Plus 88.1% Mar 2026 Official leaderboard
46 GLM-5.1 88% Apr 2026 Official leaderboard
47 Qwen3.8-2.4T-A95B 88% Aug 2026 Official leaderboard
48 o3 87.9% Apr 2025 Official leaderboard
49 Claude Opus 4.6 87.9% Feb 2026 Official leaderboard
50 GPT-5 87.8% medium Aug 2025 Official leaderboard
51 GPT-5.1 87.8% medium Nov 2025 Official leaderboard
52 Kimi K2.5 87.8% Jan 2026 Official leaderboard
53 Kimi K3 87.8% Jul 2026 Official leaderboard
54 DeepSeek V4.1 Flash 87.8% Sep 2026 Official leaderboard
55 GLM-5 87.7% Feb 2026 Official leaderboard
56 Qwen3.5-27B 87.7% Feb 2026 Official leaderboard
57 Inkling 87.7% Jul 2026 Official leaderboard
58 GPT-5.2 87.5% medium Dec 2025 Official leaderboard
59 Grok 4.20 87.5% reasoning Feb 2026 Official leaderboard
60 Qwen3.6 27B 87.5% Apr 2026 Official leaderboard
61 Claude Sonnet 4.6 87.4% Feb 2026 Official leaderboard
62 Qwen3.5-122B-A10B 87.3% Feb 2026 Official leaderboard
63 GPT-5.6 Terra 87.2% Jul 2026 Official leaderboard
64 DeepSeek V4.1 Flash 1017 87.2% Oct 2026 Official leaderboard
65 Gemma 4 31B IT 87% Apr 2026 Official leaderboard
66 DeepSeek-V4-Flash 87% Apr 2026 Official leaderboard
67 GLM-5.3 87% Aug 2026 Official leaderboard
68 Claude Opus 4.5 86.9% Nov 2025 Official leaderboard
69 Qwen 3.8 27B 86.8% Aug 2026 Official leaderboard
70 Qwen3.5-Flash 86.5% Feb 2026 Official leaderboard
71 o1 86.4% medium Dec 2024 Official leaderboard
72 Claude Opus 4 86.3% May 2025 Official leaderboard
73 GLM-4.5 86.3% Aug 2025 Official leaderboard
74 Grok 4 Fast 86.3% Sep 2025 Official leaderboard
75 Gemini 2.5 Flash Sep 2025 86.3% Sep 2025 Official leaderboard
76 Nemotron 3 Ultra 86.2% Jun 2026 Official leaderboard
77 GLM-5.2 85.9% Jun 2026 Official leaderboard
78 Claude Haiku 5.5 85.9% high Oct 2026 Official leaderboard
79 Qwen3-235B-A22B-Thinking Jul 2025 85.7% Jul 2025 Official leaderboard
80 Doubao-Seed-1.6 85.3% Jun 2025 Official leaderboard
81 Gemini 3.1 Flash-Lite 85.3% Mar 2026 Official leaderboard
82 MiMo-V2.6 Pro 85.3% Sep 2026 Official leaderboard
83 GPT-4.5 85.2% Feb 2025 Official leaderboard
84 Mistral Large 4 85.2% Oct 2026 Official leaderboard
85 Grok 4.1 Fast 85.1% Nov 2025 Official leaderboard
86 Qwen 3.6 35B-A3B 85.1% Apr 2026 Official leaderboard
87 Inkling-Small 85.1% Jul 2026 Official leaderboard
88 Kimi K2 Thinking 84.9% Nov 2025 Official leaderboard
89 Claude Sonnet 4 84.8% May 2025 Official leaderboard
90 Qwen3.7 Flash 84.8% Jul 2026 Official leaderboard
91 GLM-4.7 84.7% Dec 2025 Official leaderboard
92 GPT-5.4 Mini 84.7% medium Mar 2026 Official leaderboard
93 MiniMax-M3 84.7% Jun 2026 Official leaderboard
94 Qwen3.5-35B-A3B 84.5% Feb 2026 Official leaderboard
95 MiMo-V2-Pro 84.1% Mar 2026 Official leaderboard
96 MiMo-V2.5-Pro 83.9% Apr 2026 Official leaderboard
97 Qwen 3.6 Flash 83.9% Apr 2026 Official leaderboard
98 Muse Glimmer 83.9% Aug 2026 Official leaderboard
99 GPT-5 mini 83.3% medium Aug 2025 Official leaderboard
100 Gemini 3.5 Flash-Lite 83.2% Jul 2026 Official leaderboard
101 MiMo-V2.6 Flash 83% Sep 2026 Official leaderboard
102 GLM-4.6 82.9% Sep 2025 Official leaderboard
103 Gemma 4 26B A4B 82.9% Apr 2026 Official leaderboard
104 MiMo-V2.5 82.9% Apr 2026 Official leaderboard
105 MiniMax-M2.7 82.8% Mar 2026 Official leaderboard
106 Qwen3-235B-A22B 82.4% Apr 2025 Official leaderboard
107 Qwen3-32B 82.4% Apr 2025 Official leaderboard
108 GPT-4.1 82.3% Apr 2025 Official leaderboard
109 GPT-5.6 Luna 82.3% Jul 2026 Official leaderboard
110 Mistral Medium 3.1 82.2% Aug 2025 Official leaderboard
111 Kimi K2 Sep 2025 82.2% Sep 2025 Official leaderboard
112 o4-mini 82% Apr 2025 Official leaderboard
113 Grok 3 81.8% Apr 2025 Official leaderboard
114 Mistral Medium 3.5 81.8% Apr 2026 Official leaderboard
115 Llama 4 Maverick 81.7% Apr 2025 Official leaderboard
116 Gemma 4 12B 81.7% Jun 2026 Official leaderboard
117 Qwen3.5-9B 81.4% Feb 2026 Official leaderboard
118 Mistral Large 3 81.2% Dec 2025 Official leaderboard
119 MiniMax-M3.1-Flash 81% Oct 2026 Official leaderboard
120 Doubao-1.5-pro 80.8% Jan 2025 Official leaderboard
121 Grok-3 mini 80.7% Apr 2025 Official leaderboard
122 Grok-2 Dec 2024 80.2% Dec 2024 Official leaderboard
123 Mistral Medium 3 80% May 2025 Official leaderboard
124 Gemini 2.0 Flash Feb 2025 79.9% Feb 2025 Official leaderboard
125 Qwen3-30B-A3B 79.8% Apr 2025 Official leaderboard
126 Claude Haiku 4.5 79.5% Oct 2025 Official leaderboard
127 DeepSeek-V3.2 79.5% Dec 2025 Official leaderboard
128 o3-mini 79.4% Jan 2025 Official leaderboard
129 MiniMax-M2.1 79.1% Dec 2025 Official leaderboard
130 Llama 3.1-405B 78.8% Jul 2024 Official leaderboard
131 Nemotron 3.5 Lightning 30B-A3B 78.8% Aug 2026 Official leaderboard
132 Qwen3-14B 78.5% Apr 2025 Official leaderboard
133 DeepSeek-V3 Mar 2025 78.3% Mar 2025 Official leaderboard
134 GPT-4o Mar 2025 78.3% Mar 2025 Official leaderboard
135 DeepSeek-V3.1 78% Aug 2025 Official leaderboard
136 Gemini 2.0 Flash-Lite 77.6% Feb 2025 Official leaderboard
137 Gemini 2.5 Flash-Lite Sep 2025 77.6% Sep 2025 Official leaderboard
138 Claude 3 Opus 77% Feb 2024 Official leaderboard
139 Qwen2.5-32B 76.9% Sep 2024 Official leaderboard
140 Claude 3.7 Sonnet 76.8% Feb 2025 Official leaderboard
141 Llama 3.2 90B 76.3% Sep 2024 Official leaderboard
142 Olmo 3.1 32B Think 76% Dec 2025 Official leaderboard
143 Command A+ 75.9% May 2026 Official leaderboard
144 Qwen2.5-14B 75.6% Sep 2024 Official leaderboard
145 GPT-5.4 Nano 75.6% medium Mar 2026 Official leaderboard
146 Qwen2.5-72B 75.4% Sep 2024 Official leaderboard
147 Mistral Small 3.2 75.4% Jun 2025 Official leaderboard
148 Llama 3.3 70B 75.2% Dec 2024 Official leaderboard
149 gpt-oss-120b 75.2% Aug 2025 Official leaderboard
150 GPT-4.1 mini 75.1% Apr 2025 Official leaderboard
151 Magistral Medium 1.0 75.1% Jun 2025 Official leaderboard
152 Llama 3.1-70B 75% Jul 2024 Official leaderboard
153 GLM-4.5-Air 74.9% Jul 2025 Official leaderboard
154 Llama 4 Scout 74.5% Apr 2025 Official leaderboard
155 Gemma 4 E4B 74.4% Mar 2026 Official leaderboard
156 GPT-4 Turbo Apr 2024 74.3% Apr 2024 Official leaderboard
157 GPT-4 Jun 2023 74.1% Jun 2023 Official leaderboard
158 glm-4.7-flash 74% Jan 2026 Official leaderboard
159 Gemini 1.5 Pro Sep 2024 73.9% Sep 2024 Official leaderboard
160 gpt-oss-20b 73.9% Aug 2025 Official leaderboard
161 Nemotron 3 Nano 30B-A3B 73.8% Dec 2025 Official leaderboard
162 Qwen3-8B 73.5% Apr 2025 Official leaderboard
163 o1-mini 72.6% Sep 2024 Official leaderboard
164 Gemma 3 27B 72.6% Mar 2025 Official leaderboard
165 Mistral Small 4 72.6% Mar 2026 Official leaderboard
166 GPT-5 nano 72.4% medium Aug 2025 Official leaderboard
167 Gemini 1.5 Flash Sep 2024 72.2% Aug 2024 Official leaderboard
168 Phi-4 72.2% Dec 2024 Official leaderboard
169 LFM2.5-24B-A2B 72% Oct 2026 Official leaderboard
170 Yi-Lightning 71.8% Dec 2024 Official leaderboard
171 Yi-Large 71.5% May 2024 Official leaderboard
172 Ministral 3 14B 71.5% Dec 2025 Official leaderboard
173 Gemma 2 27B 71.2% Jun 2024 Official leaderboard
174 Granite 4.2 8B 71.1% Aug 2026 Official leaderboard
175 sonar 70.8% Jan 2025 Official leaderboard
176 Mistral Small 3 70.4% Jan 2025 Official leaderboard
177 Qwen3-4B 70.4% Apr 2025 Official leaderboard
178 Nova 2 Lite 69.6% Dec 2025 Official leaderboard
179 GPT-4o mini 69.2% Jul 2024 Official leaderboard
180 Gemma 3 12B 67.6% Mar 2025 Official leaderboard
181 MiMo-V2-Flash 67.6% Dec 2025 Official leaderboard
182 Qwen2.5-7B 66.7% Sep 2024 Official leaderboard
183 Mixtral 8x22B 65.8% Apr 2024 Official leaderboard
184 Gemma 3n E4B 64% Jun 2025 Official leaderboard
185 GPT-4.1 nano 63.5% Apr 2025 Official leaderboard
186 Claude 3.5 Haiku 63.3% Oct 2024 Official leaderboard
187 LFM3-1.7B 62.7% Oct 2026 Official leaderboard
188 Gemma 4 E2B 62.6% Mar 2026 Official leaderboard
189 Mixtral 8x7B 59.8% Dec 2023 Official leaderboard
190 Gemini 1.5 Flash 8B 59.2% Oct 2024 Official leaderboard
191 Llama 3.1-8B 59.1% Jul 2024 Official leaderboard
192 Llama 3.2 11B 59.1% Sep 2024 Official leaderboard
193 OLMo 2 32B Instruct 58.5% Mar 2025 Official leaderboard
194 Qwen3-1.7B 57.8% Apr 2025 Official leaderboard
195 Gemma 3n E2B 57.8% Jun 2025 Official leaderboard
196 Claude 3 Haiku 57% Mar 2024 Official leaderboard
197 Mistral NeMo 57% Jul 2024 Official leaderboard
198 Ministral 8B 56.4% Oct 2024 Official leaderboard
199 Phi-3.5-mini 56.1% Aug 2024 Official leaderboard
200 Phi-4-multimodal 53.2% Feb 2025 Official leaderboard
201 Ministral 3B 52.4% Oct 2024 Official leaderboard
202 Gemma 3 4B 52.4% Mar 2025 Official leaderboard
203 GPT-3.5 Turbo Jan 2024 52.3% Jan 2024 Official leaderboard
204 LFM2-8B-A1B 52.3% Oct 2025 Official leaderboard
205 Phi-3.5-vision-instruct 51.1% Aug 2024 Official leaderboard
206 Command R7B 46% Dec 2024 Official leaderboard
207 LFM2.5-1.2B-Thinking 42.8% Jan 2026 Official leaderboard
208 Qwen3-0.6B 42.7% Apr 2025 Official leaderboard
209 LFM2-1.2B 41.4% Jul 2025 Official leaderboard
210 Gemma 3 1B 33.6% Mar 2025 Official leaderboard
211 Llama 3.2 1B 28% Sep 2024 Official leaderboard
212 Llama 3.2 3B 27.6% Sep 2024 Official leaderboard
213 SmolLM2-1.7B 27.6% Oct 2024 Official leaderboard
214 SmolLM2-360M 25.9% Oct 2024 Official leaderboard
215 OLMo 2 1B Instruct 24.3% Apr 2025 Official leaderboard
216 SmolLM2-135M 21.1% Oct 2024 Official leaderboard

Rankings by part

We read 6 parts from the benchmark’s own leaderboard in 2 groups: subject and level. Each ranks models on one part of the test. Gemini 4 Argon leads 3 of them, and 3 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

Science

216 models we track have a score here.

  1. 1 Gemini 4 Argon 97.3%
  2. 2 Grok 4.6 ↑8 96.7%
  3. 3 Qwen3.5-Plus ↑31 95.1%
  4. 4 GPT-6 Astra ↓2 95.1%
  5. 5 Gemini 3 Pro ↑10 94.5%
  6. 6 Gemini 3.1 Pro ↑3 94.5%
  7. 7 Qwen 3.6 Max (Preview) ↑13 94.5%
  8. 8 Qwen3.8 Max (0902) ↑9 94.5%
  9. 9 Gemini 3.8 Flash ↓3 94.5%
  10. 10 GPT-6.1 Sol ↓2 94.5%

Scores at each effort level

Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.

Lowest effort to highest, light to dark
Gemini 4 Argon
low
92.9%
GPT-6 Astra
medium
92.7%
GPT-6.1 Terra
low
92.5%
Claude Opus 5.5
medium
91.9%
GPT-6.1 Sol
medium
91.8%
Muse Wonder
low
89.7%
Claude Sonnet 4.5
thinking
89.3%
GPT-5.4
medium
88.7%
Claude Sonnet 5.5
low
88.3%
Grok 4.20
reasoning
87.5%
Claude Haiku 5.5
low high
85.9%
GPT-5.4 Mini
medium
84.7%

Source: Official Pedagogy Benchmark leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Gemini 4 Argon high 92.9%
GPT-6 Astra high 92.7%
GPT-6.1 Terra medium 92.5%
GPT-6 Astra medium 92.3%
Claude Fable 5.1 92.2%
GPT-5.5 92.1%
Gemini 3.8 Flash 91.9%
Claude Opus 5.5 medium 91.9%
Gemini 4 Argon medium 91.9%
Claude Opus 5.5 high 91.8%
GPT-6.1 Sol medium 91.8%
GPT-6.1 Sol high 91.8%
Gemini 3.1 Pro 91.5%
Grok 4.6 91.5%
GPT-6.1 Terra low 91.4%
Gemini 4 Argon low 91.3%
GPT-5.6 Sol 91.2%
Claude Opus 5 91.1%
Gemini 3.6 Flash 90.9%
GPT-6.1 Terra high 90.9%
Gemini 3.7 Flash 90.8%
Gemini 3 Pro 90.5%
Claude Fable 5 90.5%
Qwen3.8 Max (0902) 90.3%
Qwen3.7-Max 90.1%
Muse Spark 1.1 90.1%
Qwen 3.6 Max (Preview) 89.7%
Grok 4.5 89.7%
Muse Wonder low 89.7%
Gemini 3 Flash 89.4%
Kimi K2.6 89.4%
Muse Spark 1.2 89.4%
Grok 4.7 low 89.4%
Claude Sonnet 4.5 thinking 89.3%
Tencent Hy4 preview 89.3%
Muse Wonder medium 89.3%
Claude Opus 4.8 89.2%
Qwen 3.8 Max 89.2%
Muse Wonder high 89.2%
DeepSeek-V4-Pro 89.1%
Claude Opus 4.7 89%
Muse Spark 1.3 89%
Qwen3.5-Plus 88.9%
Gemini 2.5 Pro Jun 2025 88.8%
Qwen3.8 Flash 88.8%
GPT-5.4 medium 88.7%
Grok 4.3 88.7%
Grok 4 88.4%
Gemini 3.5 Flash 88.4%
Kimi K3.1 Pro 88.4%
Claude Sonnet 5.5 high 88.3%
Claude Sonnet 5 88.2%
Qwen3.5 397B-A17B 88.1%
Qwen 3.6 Plus 88.1%
GLM-5.1 88%
Qwen3.8-2.4T-A95B 88%
o3 87.9%
Claude Opus 4.6 87.9%
GPT-5 medium 87.8%
GPT-5.1 medium 87.8%
Kimi K2.5 87.8%
Kimi K3 87.8%
DeepSeek V4.1 Flash 87.8%
Claude Sonnet 5.5 low 87.8%
GLM-5 87.7%
Qwen3.5-27B 87.7%
Inkling 87.7%
GPT-5.2 medium 87.5%
Grok 4.20 reasoning 87.5%
Qwen3.6 27B 87.5%
Claude Sonnet 5.5 medium 87.5%
Claude Sonnet 4.6 87.4%
Qwen3.5-122B-A10B 87.3%
DeepSeek-V4-Pro 87.3%
GPT-5.6 Terra 87.2%
DeepSeek V4.1 Flash 1017 87.2%
Gemma 4 31B IT 87%
DeepSeek-V4-Flash 87%
GLM-5.3 87%
Claude Opus 4.5 86.9%
Qwen 3.8 27B 86.8%
Qwen3.5-Flash 86.5%
o1 medium 86.4%
Claude Opus 4 86.3%
GLM-4.5 86.3%
Grok 4 Fast 86.3%
Gemini 2.5 Flash Sep 2025 86.3%
Nemotron 3 Ultra 86.2%
GLM-5.2 85.9%
Claude Haiku 5.5 high 85.9%
Qwen3-235B-A22B-Thinking Jul 2025 85.7%
Gemini 2.5 Flash Sep 2025 85.5%
Doubao-Seed-1.6 85.3%
Gemini 3.1 Flash-Lite 85.3%
MiMo-V2.6 Pro 85.3%
GPT-4.5 85.2%
GPT-5.4 85.2%
Mistral Large 4 85.2%
Claude Haiku 5.5 medium 85.2%
Grok 4.1 Fast 85.1%
Qwen 3.6 35B-A3B 85.1%
Inkling-Small 85.1%
Kimi K2 Thinking 84.9%
Claude Sonnet 4 84.8%
Qwen3.7 Flash 84.8%
GLM-4.7 84.7%
Gemini 3.1 Flash-Lite 84.7%
GPT-5.4 Mini medium 84.7%
MiniMax-M3 84.7%
Qwen3.5-35B-A3B 84.5%
MiMo-V2-Pro 84.1%
MiMo-V2.5-Pro 83.9%
Qwen 3.6 Flash 83.9%
Muse Glimmer 83.9%
Claude Haiku 5.5 low 83.9%
GPT-5 mini medium 83.3%
Gemini 3.5 Flash-Lite 83.2%
MiMo-V2.6 Flash 83%
GLM-4.6 82.9%
Gemma 4 26B A4B 82.9%
MiMo-V2.5 82.9%
MiniMax-M2.7 82.8%
Qwen3-235B-A22B 82.4%
Qwen3-32B 82.4%
Claude Sonnet 4.5 82.4%
GPT-4.1 82.3%
GPT-5.6 Luna 82.3%
Mistral Medium 3.1 82.2%
Kimi K2 Sep 2025 82.2%
o4-mini 82%
Grok 3 81.8%
Mistral Medium 3.5 81.8%
Llama 4 Maverick 81.7%
Gemma 4 12B 81.7%
Qwen3.5-9B 81.4%
Mistral Large 3 81.2%
MiniMax-M3.1-Flash 81%
Doubao-1.5-pro 80.8%
Grok-3 mini 80.7%
Grok-2 Dec 2024 80.2%
Mistral Medium 3 80%
Gemini 2.0 Flash Feb 2025 79.9%
Qwen3-30B-A3B 79.8%
Claude Haiku 4.5 79.5%
DeepSeek-V3.2 79.5%
o3-mini 79.4%
MiniMax-M2.1 79.1%
Llama 3.1-405B 78.8%
Nemotron 3.5 Lightning 30B-A3B 78.8%
Qwen3-14B 78.5%
DeepSeek-V3 Mar 2025 78.3%
GPT-4o Mar 2025 78.3%
DeepSeek-V3.1 78%
Gemini 2.0 Flash-Lite 77.6%
Gemini 2.5 Flash-Lite Sep 2025 77.6%
DeepSeek-V3 Mar 2025 77.1%
Claude 3 Opus 77%
Qwen2.5-32B 76.9%
Claude 3.7 Sonnet 76.8%
Llama 3.2 90B 76.3%
Gemini 2.5 Flash-Lite Sep 2025 76.2%
Olmo 3.1 32B Think 76%
DeepSeek-V4-Flash 76%
Command A+ 75.9%
Qwen2.5-14B 75.6%
GPT-5.4 Nano medium 75.6%
GPT-5.4 Mini 75.5%
Qwen2.5-72B 75.4%
Mistral Small 3.2 75.4%
Llama 3.3 70B 75.2%
gpt-oss-120b 75.2%
GPT-4.1 mini 75.1%
Magistral Medium 1.0 75.1%
Llama 3.1-70B 75%
GLM-4.5-Air 74.9%
Llama 4 Scout 74.5%
Gemma 4 E4B 74.4%
GPT-4 Turbo Apr 2024 74.3%
GPT-4 Jun 2023 74.1%
glm-4.7-flash 74%
Gemini 1.5 Pro Sep 2024 73.9%
gpt-oss-20b 73.9%
Nemotron 3 Nano 30B-A3B 73.8%
Qwen3-8B 73.5%
Grok 4.20 73.4%
o1-mini 72.6%
Gemma 3 27B 72.6%
Mistral Small 4 72.6%
GPT-5 nano medium 72.4%
Gemini 1.5 Flash Sep 2024 72.2%
Phi-4 72.2%
LFM2.5-24B-A2B 72%
Yi-Lightning 71.8%
Yi-Large 71.5%
Ministral 3 14B 71.5%
Gemma 2 27B 71.2%
Granite 4.2 8B 71.1%
sonar 70.8%
Mistral Small 3 70.4%
Qwen3-4B 70.4%
Nova 2 Lite 69.6%
GPT-4o mini 69.2%
Gemma 3 12B 67.6%
MiMo-V2-Flash 67.6%
Qwen2.5-7B 66.7%
Mixtral 8x22B 65.8%
Gemma 3n E4B 64%
GPT-4.1 nano 63.5%
Claude 3.5 Haiku 63.3%
LFM3-1.7B 62.7%
Gemma 4 E2B 62.6%
GPT-5.4 Nano 60.4%
Mixtral 8x7B 59.8%
Gemini 1.5 Flash 8B 59.2%
Llama 3.1-8B 59.1%
Llama 3.2 11B 59.1%
OLMo 2 32B Instruct 58.5%
Qwen3-1.7B 57.8%
Gemma 3n E2B 57.8%
Claude 3 Haiku 57%
Mistral NeMo 57%
Ministral 8B 56.4%
Phi-3.5-mini 56.1%
Phi-4-multimodal 53.2%
Ministral 3B 52.4%
Gemma 3 4B 52.4%
GPT-3.5 Turbo Jan 2024 52.3%
LFM2-8B-A1B 52.3%
Phi-3.5-vision-instruct 51.1%
Command R7B 46%
LFM2.5-1.2B-Thinking 42.8%
Qwen3-0.6B 42.7%
LFM2-1.2B 41.4%
Gemma 3 1B 33.6%
Llama 3.2 1B 28%
Llama 3.2 3B 27.6%
SmolLM2-1.7B 27.6%
SmolLM2-360M 25.9%
OLMo 2 1B Instruct 24.3%
SmolLM2-135M 21.1%

Scores at each effort level

Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.

Lowest effort to highest, light to dark
Gemini 4 Argon
low
92.9%
GPT-6 Astra
medium
92.7%
GPT-6.1 Terra
low
92.5%
Claude Opus 5.5
medium
91.9%
GPT-6.1 Sol
medium
91.8%
Muse Wonder
low
89.7%
Claude Sonnet 4.5
thinking
89.3%
GPT-5.4
medium
88.7%
Claude Sonnet 5.5
low
88.3%
Grok 4.20
reasoning
87.5%
Claude Haiku 5.5
low high
85.9%
GPT-5.4 Mini
medium
84.7%

Source: Official Pedagogy Benchmark leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Gemini 4 Argon high 92.9%
GPT-6 Astra high 92.7%
GPT-6.1 Terra medium 92.5%
GPT-6 Astra medium 92.3%
Claude Fable 5.1 92.2%
GPT-5.5 92.1%
Gemini 3.8 Flash 91.9%
Claude Opus 5.5 medium 91.9%
Gemini 4 Argon medium 91.9%
Claude Opus 5.5 high 91.8%
GPT-6.1 Sol medium 91.8%
GPT-6.1 Sol high 91.8%
Gemini 3.1 Pro 91.5%
Grok 4.6 91.5%
GPT-6.1 Terra low 91.4%
Gemini 4 Argon low 91.3%
GPT-5.6 Sol 91.2%
Claude Opus 5 91.1%
Gemini 3.6 Flash 90.9%
GPT-6.1 Terra high 90.9%
Gemini 3.7 Flash 90.8%
Gemini 3 Pro 90.5%
Claude Fable 5 90.5%
Qwen3.8 Max (0902) 90.3%
Qwen3.7-Max 90.1%
Muse Spark 1.1 90.1%
Qwen 3.6 Max (Preview) 89.7%
Grok 4.5 89.7%
Muse Wonder low 89.7%
Gemini 3 Flash 89.4%
Kimi K2.6 89.4%
Muse Spark 1.2 89.4%
Grok 4.7 low 89.4%
Claude Sonnet 4.5 thinking 89.3%
Tencent Hy4 preview 89.3%
Muse Wonder medium 89.3%
Claude Opus 4.8 89.2%
Qwen 3.8 Max 89.2%
Muse Wonder high 89.2%
DeepSeek-V4-Pro 89.1%
Claude Opus 4.7 89%
Muse Spark 1.3 89%
Qwen3.5-Plus 88.9%
Gemini 2.5 Pro Jun 2025 88.8%
Qwen3.8 Flash 88.8%
GPT-5.4 medium 88.7%
Grok 4.3 88.7%
Grok 4 88.4%
Gemini 3.5 Flash 88.4%
Kimi K3.1 Pro 88.4%
Claude Sonnet 5.5 high 88.3%
Claude Sonnet 5 88.2%
Qwen3.5 397B-A17B 88.1%
Qwen 3.6 Plus 88.1%
GLM-5.1 88%
Qwen3.8-2.4T-A95B 88%
o3 87.9%
Claude Opus 4.6 87.9%
GPT-5 medium 87.8%
GPT-5.1 medium 87.8%
Kimi K2.5 87.8%
Kimi K3 87.8%
DeepSeek V4.1 Flash 87.8%
Claude Sonnet 5.5 low 87.8%
GLM-5 87.7%
Qwen3.5-27B 87.7%
Inkling 87.7%
GPT-5.2 medium 87.5%
Grok 4.20 reasoning 87.5%
Qwen3.6 27B 87.5%
Claude Sonnet 5.5 medium 87.5%
Claude Sonnet 4.6 87.4%
Qwen3.5-122B-A10B 87.3%
DeepSeek-V4-Pro 87.3%
GPT-5.6 Terra 87.2%
DeepSeek V4.1 Flash 1017 87.2%
Gemma 4 31B IT 87%
DeepSeek-V4-Flash 87%
GLM-5.3 87%
Claude Opus 4.5 86.9%
Qwen 3.8 27B 86.8%
Qwen3.5-Flash 86.5%
o1 medium 86.4%
Claude Opus 4 86.3%
GLM-4.5 86.3%
Grok 4 Fast 86.3%
Gemini 2.5 Flash Sep 2025 86.3%
Nemotron 3 Ultra 86.2%
GLM-5.2 85.9%
Claude Haiku 5.5 high 85.9%
Qwen3-235B-A22B-Thinking Jul 2025 85.7%
Gemini 2.5 Flash Sep 2025 85.5%
Doubao-Seed-1.6 85.3%
Gemini 3.1 Flash-Lite 85.3%
MiMo-V2.6 Pro 85.3%
GPT-4.5 85.2%
GPT-5.4 85.2%
Mistral Large 4 85.2%
Claude Haiku 5.5 medium 85.2%
Grok 4.1 Fast 85.1%
Qwen 3.6 35B-A3B 85.1%
Inkling-Small 85.1%
Kimi K2 Thinking 84.9%
Claude Sonnet 4 84.8%
Qwen3.7 Flash 84.8%
GLM-4.7 84.7%
Gemini 3.1 Flash-Lite 84.7%
GPT-5.4 Mini medium 84.7%
MiniMax-M3 84.7%
Qwen3.5-35B-A3B 84.5%
MiMo-V2-Pro 84.1%
MiMo-V2.5-Pro 83.9%
Qwen 3.6 Flash 83.9%
Muse Glimmer 83.9%
Claude Haiku 5.5 low 83.9%
GPT-5 mini medium 83.3%
Gemini 3.5 Flash-Lite 83.2%
MiMo-V2.6 Flash 83%
GLM-4.6 82.9%
Gemma 4 26B A4B 82.9%
MiMo-V2.5 82.9%
MiniMax-M2.7 82.8%
Qwen3-235B-A22B 82.4%
Qwen3-32B 82.4%
Claude Sonnet 4.5 82.4%
GPT-4.1 82.3%
GPT-5.6 Luna 82.3%
Mistral Medium 3.1 82.2%
Kimi K2 Sep 2025 82.2%
o4-mini 82%
Grok 3 81.8%
Mistral Medium 3.5 81.8%
Llama 4 Maverick 81.7%
Gemma 4 12B 81.7%
Qwen3.5-9B 81.4%
Mistral Large 3 81.2%
MiniMax-M3.1-Flash 81%
Doubao-1.5-pro 80.8%
Grok-3 mini 80.7%
Grok-2 Dec 2024 80.2%
Mistral Medium 3 80%
Gemini 2.0 Flash Feb 2025 79.9%
Qwen3-30B-A3B 79.8%
Claude Haiku 4.5 79.5%
DeepSeek-V3.2 79.5%
o3-mini 79.4%
MiniMax-M2.1 79.1%
Llama 3.1-405B 78.8%
Nemotron 3.5 Lightning 30B-A3B 78.8%
Qwen3-14B 78.5%
DeepSeek-V3 Mar 2025 78.3%
GPT-4o Mar 2025 78.3%
DeepSeek-V3.1 78%
Gemini 2.0 Flash-Lite 77.6%
Gemini 2.5 Flash-Lite Sep 2025 77.6%
DeepSeek-V3 Mar 2025 77.1%
Claude 3 Opus 77%
Qwen2.5-32B 76.9%
Claude 3.7 Sonnet 76.8%
Llama 3.2 90B 76.3%
Gemini 2.5 Flash-Lite Sep 2025 76.2%
Olmo 3.1 32B Think 76%
DeepSeek-V4-Flash 76%
Command A+ 75.9%
Qwen2.5-14B 75.6%
GPT-5.4 Nano medium 75.6%
GPT-5.4 Mini 75.5%
Qwen2.5-72B 75.4%
Mistral Small 3.2 75.4%
Llama 3.3 70B 75.2%
gpt-oss-120b 75.2%
GPT-4.1 mini 75.1%
Magistral Medium 1.0 75.1%
Llama 3.1-70B 75%
GLM-4.5-Air 74.9%
Llama 4 Scout 74.5%
Gemma 4 E4B 74.4%
GPT-4 Turbo Apr 2024 74.3%
GPT-4 Jun 2023 74.1%
glm-4.7-flash 74%
Gemini 1.5 Pro Sep 2024 73.9%
gpt-oss-20b 73.9%
Nemotron 3 Nano 30B-A3B 73.8%
Qwen3-8B 73.5%
Grok 4.20 73.4%
o1-mini 72.6%
Gemma 3 27B 72.6%
Mistral Small 4 72.6%
GPT-5 nano medium 72.4%
Gemini 1.5 Flash Sep 2024 72.2%
Phi-4 72.2%
LFM2.5-24B-A2B 72%
Yi-Lightning 71.8%
Yi-Large 71.5%
Ministral 3 14B 71.5%
Gemma 2 27B 71.2%
Granite 4.2 8B 71.1%
sonar 70.8%
Mistral Small 3 70.4%
Qwen3-4B 70.4%
Nova 2 Lite 69.6%
GPT-4o mini 69.2%
Gemma 3 12B 67.6%
MiMo-V2-Flash 67.6%
Qwen2.5-7B 66.7%
Mixtral 8x22B 65.8%
Gemma 3n E4B 64%
GPT-4.1 nano 63.5%
Claude 3.5 Haiku 63.3%
LFM3-1.7B 62.7%
Gemma 4 E2B 62.6%
GPT-5.4 Nano 60.4%
Mixtral 8x7B 59.8%
Gemini 1.5 Flash 8B 59.2%
Llama 3.1-8B 59.1%
Llama 3.2 11B 59.1%
OLMo 2 32B Instruct 58.5%
Qwen3-1.7B 57.8%
Gemma 3n E2B 57.8%
Claude 3 Haiku 57%
Mistral NeMo 57%
Ministral 8B 56.4%
Phi-3.5-mini 56.1%
Phi-4-multimodal 53.2%
Ministral 3B 52.4%
Gemma 3 4B 52.4%
GPT-3.5 Turbo Jan 2024 52.3%
LFM2-8B-A1B 52.3%
Phi-3.5-vision-instruct 51.1%
Command R7B 46%
LFM2.5-1.2B-Thinking 42.8%
Qwen3-0.6B 42.7%
LFM2-1.2B 41.4%
Gemma 3 1B 33.6%
Llama 3.2 1B 28%
Llama 3.2 3B 27.6%
SmolLM2-1.7B 27.6%
SmolLM2-360M 25.9%
OLMo 2 1B Instruct 24.3%
SmolLM2-135M 21.1%

12 models have held number one on the official leaderboard since its release in December 2024

Lab claims count from the day the card came out, independent runs from the model’s release.

Meta xAI OpenAI Google DeepMind Anthropic
Own leaderboard Passed 85% in Dec 2024, by o1

Source: Official Pedagogy Benchmark leaderboard

O1 on the official leaderboard passed 85% in December 2024

Models released before it came out already scored that high. All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
On the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% 2023 2024 2025 2026 2027 Near the ceiling (85%) Gemini 4 Argon Gemini 3.1 Pro Gemini 2.5 Pro Grok-2 Claude 3 Opus GPT-4 0% 20% 40% 60% 80% 100% 2023 2024 2025 2026 2027 Near the ceiling (85%) Gemini 4 Argon o1

Source: Official Pedagogy Benchmark leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

Versions

More in Knowledge

All 332 in Knowledge
  • MMLU-Pro In use Harder 10-option version of MMLU with reasoning-focused questions across 14 subjects. Intern-S2-397B 89.8%
  • MMLU In use 57-subject multiple-choice exam questions from elementary to professional level. Scored by accuracy. Seed1.8 92.3%
  • SuperGPQA In use 26,529 graduate-level multiple-choice questions across 285 disciplines. Accuracy. Qwen 3.6 Max (Preview) 73.9%
  • WideSearch In use 200 bilingual broad info-seeking tasks: agent collects many atomic facts into a table. Scored by success rate and row/item F1. Tencent Hy4 preview 83.9%
  • BBEH micro average In use Models solve harder replacements for each of the 23 BIG-Bench Hard tasks, 4,520 examples in total. The score is the share of all examples answered correctly. Gemma 4 31B IT 74.4%
  • HLE-Diamond In use Models answer 1,000 cleaned expert questions from Humanity's Last Exam, 500 on reasoning and 500 on knowledge, closed book. Graded on exact answers. Claude Fable 5.2 77.9%

Data audit

All 10 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (259 entries).
  • Pass Models scored 216 models we track, plus 43 other entries on its leaderboard.
  • Pass Newest score The newest is from October 2026, for LFM3-1.7B.
  • Pass Frontier models 7 of the top 10 have a score, led by GPT-6 Astra (#4).
  • Note Headroom Saturated. The best is 92.9%, past 85% since December 2024.
Leaderboard import 3 pass
  • Pass Importer We read the board on fab-ai.org: 240 rows for models we track and 43 other entries at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check