Artificial Analysis

Artificial Analysis Humanity's Last Exam

An Artificial Analysis Humanity's Last Exam score.

In index
Knowledge
Source ↗

Models185

Top result61.4%Claude Opus 5.5

Top-3 spread5.9 pts

Year2026

Result and price

Results

185 results

#ModelResult

01

Claude Opus 5.561.4%

02

Claude Fable 5.159.1%

03

Claude Fable 555.5%

04

Claude Opus 554.9%

05

GPT-6 Astra54.7%

06

GPT-5.6 Sol49.5%

07

MiMo-V2.6-Pro49.4%

08

Muse Spark 1.348.7%

08

Claude Opus 4.848.7%

10

GPT-6 Sol47.9%

10

Gemini 3.7 Flash47.9%

12

Gemini 3.8 Flash47.8%

13

Gemini 3.1 Pro47.0%

14MAKimi K346.9%

15STStep 5 Preview46.5%

16

Muse Spark 1.146.2%

17

GPT-5.545.8%

18

Muse Spark 1.245.5%

19

GPT-5.443.7%

20

Grok 4.743.1%

20

Qwen3.8 Max Preview43.1%

22

GPT-5.6 Terra42.9%

22

Grok 4.642.9%

24

Gemini 3.5 Flash42.7%

24

Grok 4.542.7%

26

GPT-5.3 Codex42.5%

27

GLM-5.342.3%

27

Claude Opus 4.7 (Adaptive)42.3%

29

Claude Sonnet 541.3%

30

GLM-5.241.1%

31

DeepSeek V4 Pro 081341.0%

32

Gemini 3.6 Flash40.8%

33

Muse Spark40.7%

34

Qwen3.7 Max40.5%

35

Claude Opus 4.6 (Adaptive)39.9%

36

Gemini 3 Pro39.7%

37

GPT-5.6 Luna39.5%

38

DeepSeek V4.1 Flash39.2%

39

MiniMax M339.0%

40

DeepSeek V4 Flash 073138.6%

41

GPT-6 Luna38.5%

42

Qwen3.8-Flash-Next38.0%

43

GPT-5.237.7%

44MAKimi K2.637.5%

45

Grok 4.337.2%

46

GPT-5.2-Codex35.7%

46

MiMo-V2.5-Pro35.7%

48

Qwen3.7 Plus35.6%

49MAKimi K2.7 Code35.0%

50APApodex 1.134.1%

50APApodex 1.1 Mini34.1%

52

Qwen3.8-27B33.9%

53TEHy333.5%

53TEHy3 Preview33.5%

55

Claude Opus 4.733.3%

55TMInkling-Small33.3%

57TMInkling31.9%

58

Qwen 3.6 Max (preview)30.8%

59MAKimi K2.5 (Reasoning)30.7%

59MAKimi K2.530.7%

61

MiMo-V2-Pro30.4%

62

GLM-5.130.1%

62

Claude Opus 4.5 Thinking30.1%

64

MiniMax M2.729.6%

64STA.X K229.6%

66

GLM-529.3%

67UPSolar Pro 429.2%

68

GPT-5.128.5%

68

GPT-5 (high)28.5%

70

Nemotron 3 Ultra28.4%

71

GPT-5.4 nano28.3%

72

GPT-5.4 mini28.1%

73

Qwen3.6 Plus27.8%

73

GLM-5-Turbo27.8%

75

GLM-4.727.4%

76

Grok 426.7%

77

GPT-5.1-Codex-Max25.7%

77

GPT-5.1-Codex25.7%

79

GPT-5 (medium)25.4%

80

Qwen3.5-122B-A10B25.2%

81

Qwen3.5-27B23.9%

82INLing 3.0 Flash23.7%

82INLing 3.0 Flash FP823.7%

84

Gemma 4 31B23.6%

85

Qwen3.6-27B23.1%

86

Gemini 2.5 Pro22.5%

87

Qwen3.6-35B-A3B22.2%

88

MiMo-V2-Omni22.1%

89

Muse Glimmer 30B22.0%

89INLing 3.0 Flash VL22.0%

91STStep 3.7 Flash21.4%

92

Qwen3.5-35B-A3B21.0%

93

Nemotron 3 Super 100B20.8%

94

o320.1%

95

Qwen3.5 397B19.8%

95

Qwen3.5 397B (Reasoning)19.8%

97

GPT-OSS 120B19.6%

98

Grok 4.1 Fast (Reasoning)19.3%

98

Gemma 4 26B A4B19.3%

100

Claude Opus 4.619.1%

100

Grok 4 Fast (Reasoning)19.1%

102

Gemini 3.5 Flash-Lite18.8%

103MCQuasar 438B18.7%

104LAK-EXAONE 2.018.6%

105

GLM-5V-Turbo17.1%

106AATrinity-Large-Thinking15.8%

106AATrinity-Large-Preview15.8%

106

DeepSeek-R115.8%

109

Gemma 4 12B15.7%

110

Gemini 3 Flash15.0%

111

DeepSeek V3.1 (Reasoning)14.3%

112LAK-Exaone13.9%

113

Mistral Medium 3.5 128B13.8%

114

Claude Sonnet 4.613.3%

115

Claude Opus 4.513.2%

116

Claude 4.1 Opus Thinking12.5%

117COCommand A+12.0%

118

Qwen3 Max11.9%

119

Nemotron 3 Nano 30B11.4%

120

DeepSeek V3.211.2%

120IBGranite 4.2 30B11.2%

122CONorth Mini Code11.1%

123

GPT-OSS 20B11.0%

123SASarvam 105B11.0%

125

Nemotron 3.5 Lightning 30B A3B NVFP410.6%

126UPSolar Pro 310.3%

127

Mistral Small 49.9%

—

Mistral Small 4 (Reasoning)9.9%

128IBGranite 4.2 8B9.7%

129INLing 3.0 Tiny9.3%

130OPMiniCPM5-2B8.9%

131

MiMo-V2-Flash8.6%

132

Grok Code Fast 18.0%

133

o3-mini7.9%

134

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

134SASarvam 30B7.5%

136MAKimi K27.4%

136

Nemotron Ultra 253B7.4%

138

o17.0%

138

GLM-4.5-Air7.0%

140LILFM2.5-8B-A1B6.9%

141CECeleris-16.8%

142

DeepSeek V3.16.7%

143IBGranite 4.2 3B6.6%

144IBGranite-4.0-H-350M6.4%

145INLing 2.6 Flash6.3%

146LILFM2.5-2.6B6.2%

147LAExaone 4.0 1.2B5.7%

148

GLM-4.65.5%

148IBGranite-4.0-350M5.5%

150

Grok 4.1 Fast5.1%

150LILFM2.5-VL-1.6B-Extract5.1%

152

GPT-4.1 mini5.0%

152LAExaone 4.0 32B5.0%

152IBGranite-4.0-H-1B5.0%

152

Phi-4 Multimodal Instruct5.0%

156

Llama 4 Maverick4.9%

157

Nemotron 3 Nano Omni 30B A3B4.8%

157

Gemma 4 E2B4.8%

159

Gemini 2.5 Flash4.7%

160

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

160

DeepSeek R1 Distill Qwen 32B4.6%

160

Gemini 1.5 Pro4.6%

163

Gemma 3 27B4.4%

164

Claude 4 Sonnet4.3%

165

GPT-4.14.2%

165

Mistral Large 34.2%

165

GPT-4o mini4.2%

165

Gemini 1.0 Pro4.2%

169

Mistral Medium 34.1%

169

Claude 3 Haiku4.1%

171

Llama 3.1 405B4.0%

172

Gemma 4 E4B3.8%

172

Llama 4 Scout3.8%

172

GPT-4.1 nano3.8%

172

Phi-43.8%

176UPSolar Pro 23.7%

177FAUltravox v0.6 Llama 3.3 70B3.6%

178

Qwen2.5 Coder 32B Instruct3.5%

179

Mistral Large 23.3%

180AMNova Pro3.2%

181

GPT-4 Turbo3.1%

182

DeepSeek V32.9%

183

Claude 3 Opus2.8%

184

GPT-4o2.4%