Inference
Vals AI

AIME

Challenging national math exam given to top high-school students

ModelsModels with a result.91
Top resultBest result on this test.98.1%Gemini 3.1 Pro Preview
Top-3 spreadPoints from first to third.1.3 pts
UpdatedDate the source changed the results.16 Apr 2026

Result and cost

020406080100$0.0001$0.0003$0.001$0.003$0.01$0.03$0.1$0.3$1$3RESULTUS DOLLARS PER TASK · LOG SCALE

Results

92 results
#ModelResultThe score from the source.CostUS dollars to run one task.TimeTime to run one task.
01Gemini 3.1 Pro Preview98.1%$0.1272 min
02GPT-5.296.9%$0.1352 min
02Muse Spark96.9%<$0.013 min
04Gemini 3 Pro96.7%$0.1692 min
05GPT-5.496.7%$0.1382 min
06Grok 4.20 (Reasoning)96.5%$0.0592 min
07Claude Opus 4.796.3%$0.4513 min
08Claude Opus 4.6 (Adaptive)95.6%$0.3445 min
08Gemini 3 Flash Preview95.6%$0.04684 s
08GPT-5.4 mini95.6%$0.1295 min
08Kimi K2.5 Thinking95.6%$0.0645 min
12Claude Opus 4.5 Thinking95.4%$1.324 min
13Qwen3.6 Plus94.6%$0.0949 min
14GPT-593.4%$0.0935 min
15GPT-5.193.3%$0.153 min
15GLM-4.793.3%$0.0439 min
17GLM-4.692.7%$0.11612 min
18GPT-OSS 120B92.6%$0.013 min
19Qwen3.5 Flash92.5%$0.0114 min
20Claude Sonnet 4.692.3%$0.4365 min
21GLM-5.191.9%$0.0716 min
21Grok 4.1 Fast (Reasoning)91.9%<$0.012 min
23GLM-5 Thinking91.7%$0.078 min
24GPT-5 mini91.5%$0.0232 min
25Grok 4 Fast (Reasoning)91.3%<$0.0135 s
26MiniMax M2.791.0%$0.0234 min
27Grok 490.6%$0.1742 min
28GPT-5.4 nano88.8%<$0.0166 s
28MiniMax M2.588.8%$0.0416 min
30Claude Sonnet 4.5 Thinking88.2%$0.2674 min
31GLM-4.586.7%$0.0476 min
32o3-mini86.5%$0.0513 min
33Qwen3.5 Plus Thinking86.0%$0.0454 min
33GPT-OSS 20B86.0%$0.0114 min
35Gemini 2.5 Pro Exp 03.2585.8%$0.0272 min
36Kimi K2 Thinking85.4%$0.04515 min
37o385.3%$0.064 min
38Grok 3 Mini Fast High Reasoning85.0%<$0.012 min
39DeepSeek V3.2 (Thinking)84.6%<$0.016 min
40Qwen3 235B A22B84.0%$0.0124 min
41o4-mini (high)83.7%$0.02655 s
42Magistral Medium83.5%$0.0877 min
43Gemini 3.1 Flash Lite Preview83.3%<$0.0141 s
44Claude Haiku 4.5 Thinking82.7%$0.1743 min
45GPT-5 nano81.2%<$0.013 min
46Qwen3 Max81.0%$0.03317 s
47Magistral Small80.7%$0.0254 min
48Claude Opus 4.1 Thinking78.2%$0.864 min
49MiniMax M2.177.9%$0.0244 min
50Claude Opus 4.576.9%$0.08917 s
51Claude Sonnet 4 Thinking76.3%$0.2955 min
52DeepSeek-R174.0%$0.0533 min
53o171.5%$0.6593 min
54Grok 3 Mini Fast Low Reasoning70.6%<$0.0131 s
55DeepSeek V3.264.8%<$0.012 min
56Kimi K2 Instruct62.7%$0.0132 min
57Qwen3 Max60.7%$0.0233 min
58Grok 3 [Beta]58.8%$0.09164 s
59Llama 3.3 Nemotron Super 49B V1 42e84561 Thinking53.5%$03 min
60Gemini 2.5 Flash Preview Thinking51.5%<$0.0162 s
61Gemini 2.5 Flash (Preview)49.8%<$0.0158 s
62GPT-4.1 mini49.4%<$0.0133 s
63Claude 3.7 Sonnet Thinking44.6%$0.3955 min
64Claude Opus 4.144.2%$030 s
65Mistral Medium42.3%$066 s
66Gemini 2.5 Flash Lite Preview Thinking42.1%<$0.0151 s
67Claude Opus 441.3%$0.1137 s
68DeepSeek V339.9%<$0.0151 s
69GPT-4.139.6%$0.053 min
70Claude Sonnet 438.5%$0.02223 s
71Grok 4 Fast Non Reasoning33.3%<$0.016 s
72Gemini 2.0 Flash 00129.8%<$0.0111 s
73Grok 4.1 Fast Non Reasoning27.1%<$0.0117 s
74GPT-4.1 nano26.5%<$0.0112 s
75Gemini 2.5 Flash Lite Preview26.3%<$0.0119 s
76Mistral Large26.0%<$0.012 min
77Llama4 Maverick Instruct Basic25.2%<$0.0116 s
78Claude 3.7 Sonnet22.3%$0.02319 s
79Llama 4 Scout 17B 16E Instruct19.0%<$0.0122 s
80Gemini 1.5 Pro 00218.8%<$0.0111 s
81Gemini 1.5 Flash 00217.3%<$0.016 s
82Llama 3.3 70B Instruct Turbo16.0%<$0.0111 s
83Grok 215.2%$0.01558 s
84Command A13.3%$0.01223 s
85GPT-4o12.9%$0.01216 s
86GPT-4o mini11.5%<$0.0129 s
87Claude 3.5 Sonnet10.0%<$0.019 s
88Llama 3.3 Nemotron Super 49B V1 42e845619.4%$016 s
89Mistral Small4.6%<$0.0113 s
90Claude 3.5 Haiku3.3%<$0.019 s
91Jamba Large 1.60.4%<$0.0119 s
91Jamba Mini 1.60.4%<$0.017 s