#ModelAccuracyPaper
1open-ended56.87link
2mcq + open-ended55.66link
3Nemotron-CrossThink (B_nd)55.66link
4orz55.20link
5Short54.50link
6Long53.30link
7Qwen2.5-Math +P-GRPO (Math+RM)51.50link
8Eurus-2-PRIME48.20link
9Qwen2.5-Math +GRPO (Math)48.00link
10β„³44.75link
11Qwen2.5-Math-Inst43.40link
12GPT-4o40.80link
13Llama-3.1-Inst (70B)38.90link
14Qwen2.5-Math24.50link
Avg avg Leaderboard