Benchgen

MGSM — Results

RankModelScore
1llama-4-maverick0.923
2o3-mini0.92
3claude-3-5-sonnet0.916
4llama-3-3-70b-instruct0.911
5llama-4-scout0.906
6gpt-4o0.905
7o10.893
8gpt-4-turbo0.885
9gemini-1-5-pro0.875
10gpt-4o-mini0.87
11llama-3-2-90b-instruct0.869
12claude-3-5-haiku0.856
13qwen3-235b-a22b0.835
14gemini-1-5-flash0.826
15phi-40.806
16phi-4-mini0.639
17phi-3-5-moe-instruct0.587
18llama-3-2-3b-instruct0.582
19phi-3-5-mini-instruct0.479

MGSM — Evaluations

1 phaseActive

No evaluations yet for this benchmark.