Artificial Analysis

Artificial Analysis Humanity's Last Exam

An Artificial Analysis Humanity's Last Exam score.

In index
Knowledge
Best in knowledge
Source ↗

Models223

Top result74.3%Claude Fable 5.2

Top-3 spread15.2 pts

Year2026

Result and price

147 of 223 results

Results

223 results

#ModelResult

01Claude Fable 5.274.3%

02Claude Opus 5.561.4%

03Claude Fable 5.159.1%

04Gemini 4 Argon57.1%

05Claude Fable 555.5%

06Claude Sonnet 5.555.0%

07Claude Opus 554.9%

08GPT-6 Astra54.7%

09Muse Wonder Full53.6%

10GPT-6.1 Sol52.9%

11Kimi K3.1 Pro52.5%

12Muse Wonder50.1%

13GPT-5.6 Sol49.5%

14MiMo-V2.6-Pro49.4%

15Muse Spark 1.348.7%

15Claude Opus 4.848.7%

17MiniMax M3.1 Flash48.0%

18GPT-6 Sol47.9%

18Gemini 3.7 Flash47.9%

20Gemini 3.8 Flash47.8%

21Gemini 3.1 Pro47.0%

22Kimi K346.9%

23Step 5 Preview46.5%

24GPT-6.1 Terra46.2%

24Muse Spark 1.146.2%

26GPT-5.545.8%

27Muse Spark 1.245.5%

28Claude Haiku 5.544.4%

29Inkling 2 Small44.1%

30Muse Spark Open43.9%

31GPT-5.443.7%

32Grok 4.743.1%

32Qwen3.8 Max Preview43.1%

34GPT-5.6 Terra42.9%

34Grok 4.642.9%

36Gemini 3.5 Flash42.7%

36Grok 4.542.7%

38GPT-5.3 Codex42.5%

39GLM-5.342.3%

39Claude Opus 4.7 (Adaptive)42.3%

41Claude Sonnet 541.3%

42GLM-5.241.1%

43DeepSeek V4 Pro 081341.0%

44Gemini 3.6 Flash40.8%

45Muse Spark40.7%

46Qwen3.7 Max40.5%

47Claude Opus 4.6 (Adaptive)39.9%

48Gemini 3 Pro39.7%

49GPT-5.6 Luna39.5%

50Ling 3.1 Flash39.4%

51DeepSeek V4.1 Flash 101739.2%

52MiniMax M339.0%

53DeepSeek V4 Flash 073138.6%

54GPT-6 Luna38.5%

55Qwen3.8-Flash-Next38.0%

56GPT-5.237.7%

57Kimi K2.637.5%

58Grok 4.337.2%

59GPT-5.2-Codex35.7%

59MiMo-V2.5-Pro35.7%

61Qwen3.7 Plus35.6%

62MiMo-V2.6-Flash35.1%

63Kimi K2.7 Code35.0%

64Beam 1.0 501B-A23B34.7%

65Apodex 1.134.1%

65Apodex 1.1 Mini34.1%

67Qwen3.8-27B33.9%

68Hy3 Preview33.5%

68Hy333.5%

70Claude Opus 4.733.3%

70Inkling-Small33.3%

72Inkling31.9%

73Qwen 3.6 Max (preview)30.8%

74Kimi K2.5 (Reasoning)30.7%

74Kimi K2.530.7%

74Nemotron 3.5 Super VL 120B A12B30.7%

74Nemotron 3.5 Super VL 120B A12B NVFP430.7%

78MiMo-V2-Pro30.4%

79Claude Opus 4.5 Thinking30.1%

79GLM-5.130.1%

81Solar Pro 429.8%

82A.X K229.6%

82MiniMax M2.729.6%

84GLM-529.3%

85Qwen3.5 397B (Reasoning)29.0%

85Qwen3.5 397B29.0%

87GPT-5.128.5%

87GPT-5 (high)28.5%

89Nemotron 3 Ultra28.4%

90GPT-5.4 nano28.3%

91GPT-5.4 mini28.1%

92Qwen3.6 Plus27.8%

92GLM-5-Turbo27.8%

94GLM-4.727.4%

95Grok 426.7%

96GPT-5.1-Codex-Max25.7%

96GPT-5.1-Codex25.7%

98GPT-5 (medium)25.4%

99Qwen3.5-122B-A10B25.2%

100Nemotron 3.5 Bolt 74B-A9B24.7%

101Qwen3.5-27B23.9%

102Ling 3.0 Flash23.7%

103Gemma 4 31B23.6%

104Qwen3.6-27B23.1%

105Gemini 2.5 Pro22.5%

106Qwen3.6-35B-A3B22.2%

107MiMo-V2-Omni22.1%

108Muse Glimmer 30B22.0%

108Ling 3.0 Flash VL22.0%

110Step 3.7 Flash21.4%

111Qwen3.5-35B-A3B21.0%

112o320.1%

113GPT-OSS 120B19.6%

114Grok 4.1 Fast (Reasoning)19.3%

114Gemma 4 26B A4B19.3%

116Claude Opus 4.619.1%

116Grok 4 Fast (Reasoning)19.1%

118Gemini 3.5 Flash-Lite18.8%

119Quasar 438B18.7%

120K-EXAONE 2.018.6%

121GLM-5V-Turbo17.1%

122o4-mini16.5%

123Trinity-Large-Thinking15.8%

123Trinity-Large-Preview15.8%

123DeepSeek-R115.8%

126Gemma 4 12B15.7%

127Gemini 3 Flash15.0%

128DeepSeek V3.1 (Reasoning)14.3%

129K-Exaone13.9%

130Mistral Medium 3.5 128B13.8%

131Claude Sonnet 4.613.3%

132Claude Opus 4.513.2%

133Claude 4.1 Opus Thinking12.5%

134Claude 4 Opus Thinking12.3%

135Command A+12.0%

136Qwen3 Max11.9%

137Mercury 2.511.8%

138DeepSeek V3.211.2%

138Granite 4.2 30B11.2%

140North Mini Code11.1%

141GPT-OSS 20B11.0%

141Sarvam 105B11.0%

143Nemotron 3.5 Lightning 30B A3B NVFP410.6%

144Solar Pro 310.3%

145Qwen3.5-4B9.9%

145Mistral Small 49.9%

—Mistral Small 4 (Reasoning)9.9%

147Magistral Medium9.8%

148Claude 3.7 Sonnet Thinking9.7%

148Granite 4.2 8B9.7%

150Ling 3.0 Tiny9.3%

151MiniCPM5-2B8.9%

152MiMo-V2-Flash8.6%

153Grok Code Fast 18.0%

154o3-mini7.9%

155Magistral Small7.6%

156Sarvam 30B7.5%

156Qwen3-Omni-30B-A3B-Thinking7.5%

158Kimi K27.4%

158QwQ-32B7.4%

158Nemotron Ultra 253B7.4%

161o17.0%

161GLM-4.5-Air7.0%

163LFM2.5-8B-A1B6.9%

164Celeris-16.8%

165DeepSeek V3.16.7%

166Granite 4.2 3B6.6%

167Granite-4.0-H-350M6.4%

168Ling 2.6 Flash6.3%

169LFM2.5-2.6B6.2%

169Claude 4 Opus6.2%

171Exaone 4.0 1.2B5.7%

172GLM-4.65.5%

172Granite-4.0-350M5.5%

174Grok 4.1 Fast5.1%

174LFM2.5-VL-1.6B-Extract5.1%

176GPT-4.1 mini5.0%

176Exaone 4.0 32B5.0%

176Granite-4.0-H-1B5.0%

176Phi-4 Multimodal Instruct5.0%

180Llama 4 Maverick4.9%

181Nemotron 3 Nano Omni 30B A3B4.8%

181Gemma 4 E2B4.8%

183DeepSeek V3 03244.7%

183Gemini 2.5 Flash4.7%

185Nemotron 3 Nano 30B4.6%

185DeepSeek R1 Distill Qwen 32B4.6%

185Gemini 1.5 Pro4.6%

185Qwen3-Omni-30B-A3B-Instruct4.6%

189Gemma 3 27B4.4%

190Claude 4 Sonnet4.3%

190Gemini 2.0 Flash4.3%

190Mistral Small 3.14.3%

190Llama 3.1 8B4.3%

194GPT-4.14.2%

194Claude 3.7 Sonnet4.2%

194Mistral Large 34.2%

194GPT-4o mini4.2%

194Gemini 1.0 Pro4.2%

199Mistral Medium 34.1%

199Claude 3 Haiku4.1%

201Llama 3.1 405B4.0%

202Devstral Small3.8%

202Gemma 4 E4B3.8%

202Phi-4-reasoning3.8%

202Llama 4 Scout3.8%

202GPT-4.1 nano3.8%

202Phi-43.8%

208Gemini 2.5 Flash-Lite3.7%

208Solar Pro 23.7%

210o1-mini3.6%

210Llama 3.3 70B3.6%

210Llama 3.1 70B3.6%

210Claude 3.5 Haiku3.6%

210Ultravox v0.6 Llama 3.3 70B3.6%

215Qwen2.5 Coder 32B Instruct3.5%

216Mistral Large 23.3%

217Nova Pro3.2%

217Gemini 1.5 Flash3.2%

219GPT-4 Turbo3.1%

220DeepSeek V32.9%

221Claude 3 Opus2.8%

222GPT-4o2.4%