Pass IndexThe State of AISign in

AIME 2025

AIME 2025 is run by MathArena (ETH Zurich SRI Lab). It has ranked 61 models, of which the catalogue holds 23, scoring from 88.33 to 100 on Accuracy (± 95% CI).

measured by MathArena (ETH Zurich SRI Lab) · the board itself

PlaceModelMetricScore
1stof 61GPT-5.2Accuracy (± 95% CI) (xhigh)100
3rdof 61Step 3.5 FlashAccuracy (± 95% CI)98.33
4thof 61Gemini 3 Flash PreviewAccuracy (± 95% CI)97.5
5thof 61GLM-5Accuracy (± 95% CI)96.67
6thof 61Kimi K2.5Accuracy (± 95% CI) (think)95.83
6thof 61DeepSeek-V3.2-SpecialeAccuracy (± 95% CI)95.83
8thof 61GPT-5Accuracy (± 95% CI) (high)95
8thof 61gemini-3-pro-previewAccuracy (± 95% CI) (preview)95
10thof 61DeepSeek v3.2Accuracy (± 95% CI) (think)94.17
10thof 61GPT-5.1Accuracy (± 95% CI) (high)94.17
12thof 61GLM-4.5Accuracy (± 95% CI)93.33
13thof 61Kimi K2 ThinkingAccuracy (± 95% CI)92.5
13thof 61Grok 4Accuracy (± 95% CI)92.5
15thof 61DeepSeek V3.2 ExpAccuracy (± 95% CI) (think)91.67
15thof 61GLM-4.6Accuracy (± 95% CI)91.67
15thof 61o4 Mini HighAccuracy (± 95% CI)91.67
18thof 61DeepSeek-V3.1Accuracy (± 95% CI) (think)90.83
20thof 61GPT OSS 120BAccuracy (± 95% CI) (high)90
21stof 61Grok 4.1 Fast ReasoningAccuracy (± 95% CI)89.17
21stof 61GPT OSS 20BAccuracy (± 95% CI) (high)89.17
21stof 61DeepSeek-R1 (0528)Accuracy (± 95% CI)89.17
21stof 61o3Accuracy (± 95% CI) (high)89.17
25thof 61Gemini 2.5 ProAccuracy (± 95% CI)88.33

Read from the board on 2026-08-25