Artificial Analysis

Artificial Analysis GPQA Diamond

An Artificial Analysis GPQA Diamond score.

In index
Knowledge
Best in knowledge
Source ↗

Models200

Top result96.1%GPT-6 Astra

Top-3 spread1.2 pts

Year2026

Result and price

128 of 200 results

Results

200 results

#ModelResult

01GPT-6 Astra96.1%

02Gemini 3.8 Flash95.3%

03Grok 4.694.9%

04Gemini 3.7 Flash94.5%

05GPT-5.6 Sol94.1%

05Gemini 3.1 Pro94.1%

07Claude Fable 5.193.7%

08Muse Spark 1.393.5%

08GPT-5.593.5%

08Kimi K393.5%

11Claude Opus 593.2%

12Grok 4.593.1%

13MiniMax M392.9%

14DeepSeek V4 Pro 081392.8%

14Qwen3.8 Max Preview92.8%

14Gemini 3.6 Flash92.8%

17Claude Fable 592.6%

18GPT-5.6 Terra92.5%

19Qwen3.8-Flash-Next92.3%

19Qwen3.7 Max92.3%

21Gemini 3.5 Flash92.2%

22Claude Opus 4.892.0%

22GPT-5.492.0%

24GLM-5.391.7%

25GPT-5.3 Codex91.5%

26Claude Opus 4.7 (Adaptive)91.4%

27GPT-5.6 Luna91.1%

27Claude Sonnet 591.1%

27Kimi K2.691.1%

30DeepSeek V4 Flash 073190.8%

30Gemini 3 Pro90.8%

32Qwen3.8-27B90.5%

33Muse Spark 1.290.4%

34GPT-5.290.3%

35Grok 4.390.1%

36Qwen3.7 Plus90.0%

37GPT-5.2-Codex89.9%

38Muse Spark 1.189.8%

39Hy3 Preview89.7%

39Hy389.7%

41Claude Opus 4.6 (Adaptive)89.6%

41Kimi K2.7 Code89.6%

43GLM-5.289.5%

43Inkling-Small89.5%

45Qwen3.5 397B (Reasoning)89.3%

45Qwen3.5 397B89.3%

47Solar Pro 489.1%

48Muse Spark Open89.0%

49Qwen 3.6 Max (preview)88.8%

50Claude Opus 4.788.5%

51Muse Spark88.4%

52Qwen3.6 Plus88.2%

53Kimi K2.5 (Reasoning)87.9%

53Kimi K2.587.9%

55Grok 487.7%

56GPT-5.4 mini87.5%

57MiniMax M2.787.4%

58GPT-5.187.3%

59Inkling87.2%

60MiMo-V2-Pro87.0%

61GLM-5.186.8%

62Nemotron 3 Ultra86.7%

63MiMo-V2.5-Pro86.6%

63Claude Opus 4.5 Thinking86.6%

65Apodex 1.186.4%

65Apodex 1.1 Mini86.4%

67Ling 3.0 Flash VL86.2%

68GPT-5.1-Codex-Max86.0%

68GPT-5.1-Codex86.0%

70GLM-4.785.9%

71Qwen3.5-27B85.8%

72A.X K285.7%

72Qwen3.5-122B-A10B85.7%

72Gemma 4 31B85.7%

75Ling 3.0 Flash85.5%

76GPT-5 (high)85.4%

77Grok 4.1 Fast (Reasoning)85.3%

78GLM-5-Turbo84.7%

78Grok 4 Fast (Reasoning)84.7%

80Qwen3.5-35B-A3B84.5%

80o3-pro84.5%

82Gemini 2.5 Pro84.4%

83GPT-5 (medium)84.2%

83Qwen3.6-27B84.2%

85Qwen3.6-35B-A3B84.1%

86Claude Opus 4.684.0%

87Gemini 3.5 Flash-Lite83.8%

88Muse Glimmer 30B83.5%

89K-EXAONE 2.082.9%

90MiMo-V2-Omni82.8%

91o382.7%

92GLM-582.0%

93GPT-5.4 nano81.7%

94DeepSeek-R181.3%

95Gemini 3 Flash81.2%

96Claude Opus 4.581.0%

97Step 3.7 Flash80.9%

97GLM-5V-Turbo80.9%

97Claude 4.1 Opus Thinking80.9%

100Claude Sonnet 4.679.9%

101Claude 4 Opus Thinking79.6%

102Gemma 4 26B A4B79.2%

103o4-mini78.4%

104K-Exaone78.3%

105GPT-OSS 120B78.2%

106DeepSeek V3.1 (Reasoning)77.9%

107Claude 3.7 Sonnet Thinking77.2%

108Qwen3.5-4B77.1%

109Mistral Small 476.9%

—Mistral Small 4 (Reasoning)76.9%

110Kimi K276.6%

111o1-preview76.5%

112Qwen3 Max76.4%

113Command A+76.1%

114North Mini Code75.7%

115Gemma 4 12B75.3%

116Trinity-Large-Thinking75.2%

116Trinity-Large-Preview75.2%

118DeepSeek V3.275.1%

119Mistral Medium 3.5 128B74.8%

119o3-mini74.8%

121o174.7%

122Nemotron 3.5 Lightning 30B A3B NVFP474.3%

123Sarvam 105B73.8%

124DeepSeek V3.173.5%

125Ling 3.0 Tiny73.4%

126GLM-4.5-Air73.3%

127Quasar 438B73.2%

128Nemotron Ultra 253B72.8%

129Grok Code Fast 172.7%

130Qwen3-Omni-30B-A3B-Thinking72.6%

131Solar Pro 372.4%

132MiniCPM5-2B70.2%

133Claude 4 Opus70.1%

134GPT-OSS 20B68.8%

135Claude 4 Sonnet68.3%

135Gemini 2.5 Flash68.3%

137Mistral Large 368.0%

138Magistral Medium67.9%

139Llama 4 Maverick67.1%

140GPT-4.166.6%

141GPT-4.1 mini66.4%

142MiMo-V2-Flash65.6%

142Claude 3.7 Sonnet65.6%

144DeepSeek V3 032465.5%

145Granite 4.2 30B64.4%

146Magistral Small64.1%

147Grok 4.1 Fast63.7%

148Sarvam 30B63.3%

149GLM-4.663.2%

150Granite 4.2 8B63.1%

150Celeris-163.1%

152Exaone 4.0 32B62.8%

153Gemini 2.0 Flash62.3%

154Qwen3-Omni-30B-A3B-Instruct62.0%

155DeepSeek R1 Distill Qwen 32B61.5%

156o1-mini60.3%

157Ling 2.6 Flash59.3%

157QwQ-32B59.3%

159Gemini 1.5 Pro58.9%

160Llama 4 Scout58.7%

161Mistral Medium 357.8%

162Gemma 4 E4B57.6%

163Phi-4-reasoning57.5%

163Phi-457.5%

165Solar Pro 256.1%

166Granite 4.2 3B55.9%

167LFM2.5-2.6B55.8%

168DeepSeek V355.7%

169GPT-4o54.3%

170Llama 3.1 405B51.5%

171LFM2.5-8B-A1B51.3%

172GPT-4.1 nano51.2%

173Nova Pro49.9%

174Llama 3.3 70B49.8%

174Ultravox v0.6 Llama 3.3 70B49.8%

176Llama 3.1 70B49.1%

177Claude 3 Opus48.9%

178Mistral Large 248.6%

179Gemini 2.5 Flash-Lite47.4%

180Nemotron 3 Nano Omni 30B A3B46.9%

181Gemini 1.5 Flash46.3%

182Mistral Small 3.145.4%

183Gemma 4 E2B43.3%

184Gemma 3 27B42.8%

185GPT-4o mini42.6%

186Exaone 4.0 1.2B42.4%

187DeepSeek-V2.542.3%

188Qwen2.5 Coder 32B Instruct41.7%

189Devstral Small41.4%

190Claude 3.5 Haiku40.8%

191Nemotron 3 Nano 30B39.9%

192Claude 3 Haiku37.4%

193Phi-4 Multimodal Instruct31.5%

194LFM2.5-VL-1.6B-Extract28.9%

195Gemini 1.0 Pro27.7%

196Llama 3.1 8B27.0%

197Granite-4.0-H-1B26.3%

198Granite-4.0-350M26.1%

199Granite-4.0-H-350M25.7%