Benchgen

AIME 2025 — Results

RankModelScore
1ripple-65b-preview98.3
2kolibri-196.9
3granite-4-2-30b89.17
4twil-8b88.3
5granite-4-2-8b86.67
6granite-4-2-3b78.33
7lfm2-5-2-6b51.87
8k2-horizon-0-9b41.7
9gemini-3-pro1
10gpt-5-2-pro-2025-12-111
11gpt-5-21
12kimi-k2-thinking-09051
13claude-opus-4-60.998
14gemini-3-flash0.997
15seed-2-0-pro0.983
16kimi-k2-50.961
17deepseek-v3-2-speciale0.96
18glm-4-70.957
19gpt-50.946
20gpt-5-1-instant0.94
21gpt-5-1-thinking0.94
22gpt-5-10.94
23deepseek-v3-2-thinking0.931
24deepseek-v3-20.931
25seed-2-0-lite0.93

AIME 2025 — Phases

1 phaseActive

Phase 1·

AIME 2025

Aug 3, 2026
Current