List your company

Open-Form Math on Math

93.7AccuracyBERT-Judge
57.50866.90476.385.696Apr 10, 2026

Evaluation results

MethodLinks
BERT-Judgeclean_name=BERT-as-a-Judge2026.0493.7
Regex2026.0473.4
LLM-Judgebackbone=Qwen-3 0.6B2026.0458.9