List your company

Math Word Problem Solving on GSM8K (Acc, AUROC, ECE)

0.9636AccuracyQwen3 30B
0.6419280.7254390.808950.892461Feb 18, 2026

Evaluation results

MethodLinks
Qwen3 30Bconfidence_type=normalized, evaluation_protocol=self-evaluation2026.020.96360.7870.026
Gemma3 12Bconfidence_type=normalized, evaluation_protocol=self-evaluation2026.020.92870.8010.045
Qwen3 4Bconfidence_type=normalized, evaluation_protocol=self-evaluation2026.020.92490.7730.069
Gemma3 4Bconfidence_type=normalized, evaluation_protocol=self-evaluation2026.020.86050.6860.121
GLM4 9Bconfidence_type=normalized, evaluation_protocol=self-evaluation2026.020.65430.7340.137