Benchgen

MATH-500 — Results

RankModelScore
1longcat-flash-thinking0.992
2glm-4-50.982
3glm-4-5-air0.981
4kimi-k2-instruct-09050.974
5kimi-k2-instruct0.974
6llama-3-1-nemotron-ultra-253b-v10.97
7longcat-flash-lite0.968
8llama-3-3-nemotron-super-49b-v10.966
9longcat-flash-chat0.964
10kimi-k1-50.962
11deepseek-r1-05280.959
12deepseek-v3-03240.94
13qwq-32b0.906
14deepseek-v30.902

MATH-500 — Leaderboard

1 phaseActive

MATH-500 Leaderboard

No self-reported entries.