List your company

Mathematical Reasoning on GSM8K (Accuracy, Genuine-Followup Rate)

96.8AccuracyQwen3.5-397B-A17B
89.20891.17993.1595.121Apr 2, 2026

Evaluation results

MethodLinks
Qwen3.5-397B-A17B2026.0496.80.8
Qwen3.5-27B2026.0495.80
gpt-oss-120b2026.0495.80
GLM-4.72026.0495.31.2
gpt-oss-20b2026.0493.91.4
Qwen3.5-9B2026.0489.50.5