List your company

Mathematical Reasoning on AIME 2025 (Pass@1, Timing, and Steps)

67.9Pass@1PS-ada
56.66859.58462.565.416May 6, 2026

Evaluation results

MethodLinks
PS-adaModel=Qwen3-8B2026.0567.936.931.319016169977.7
PS-adaModel=Qwen3-4B2026.0566.229.523.114011075780.2
BaselineModel=Qwen3-4B2026.055949.545.423021177566
BaselineModel=Qwen3-8B2026.0557.145.443.826025162861