Benchgen

AIME 2025 — Results

RankModelScore
1lfm2-5-2-6b51.87
2gemini-3-pro1
3gpt-5-2-pro-2025-12-111
4gpt-5-21
5kimi-k2-thinking-09051
6claude-opus-4-60.998
7gemini-3-flash0.997
8seed-2-0-pro0.983
9kimi-k2-50.961
10deepseek-v3-2-speciale0.96
11glm-4-70.957
12gpt-50.946
13gpt-5-1-instant0.94
14gpt-5-1-thinking0.94
15gpt-5-10.94
16deepseek-v3-2-thinking0.931
17deepseek-v3-20.931
18seed-2-0-lite0.93
19o4-mini0.927
20gpt-oss-120b-high0.925
21qwen3-235b-a22b-thinking-25070.923
22grok-40.917
23gpt-5-mini0.911
24grok-3-mini0.908
25qwen3-vl-235b-a22b-thinking0.897

AIME 2025 — Phases

1 phaseActive

Phase 1·

AIME 2025

Aug 3, 2026
Current