List your company

Mathematical Reasoning on Avg. (held out)

62.8AccuracySC @ 64
20.1631.2342.353.37May 8, 2026

Evaluation results

MethodLinks
SC @ 64Base Model=Qwen3-8B, Type=Handcrafted2026.0562.81,195.7
ASCBase Model=Qwen3-8B, Type=Handcrafted2026.0562.8485.7
ESCBase Model=Qwen3-8B, Type=Handcrafted2026.0562.8927.6
AutoTTS (β=1.0)Base Model=Qwen3-8B, Type=Discovered2026.0562.7710.8
Parallel-ProbeBase Model=Qwen3-8B, Type=Handcrafted2026.0562872
AutoTTS (β=0.5)Base Model=Qwen3-8B, Type=Discovered2026.0561.1379
AutoTTS (β=1.0)Base Model=Qwen3-4B, Type=Discovered2026.0560.5648.6
Parallel-ProbeBase Model=Qwen3-4B, Type=Handcrafted2026.0560.4839.2
ASCBase Model=Qwen3-4B, Type=Handcrafted2026.0560.2333.1
SC @ 64Base Model=Qwen3-4B, Type=Handcrafted2026.0560.11,128.2
ESCBase Model=Qwen3-4B, Type=Handcrafted2026.0560.1891.8
AutoTTS (β=0.5)Base Model=Qwen3-4B, Type=Discovered2026.0559.8348.7
AutoTTS (β=1.0)Base Model=Qwen3-1.7B, Type=Discovered2026.0540.6646.1
AutoTTS (β=0.5)Base Model=Qwen3-1.7B, Type=Discovered2026.0538.6343.5
SC @ 64Base Model=Qwen3-1.7B, Type=Handcrafted2026.0534.31,093.5
ASCBase Model=Qwen3-1.7B, Type=Handcrafted2026.0534.3593.6
ESCBase Model=Qwen3-1.7B, Type=Handcrafted2026.0534.3964.3
Parallel-ProbeBase Model=Qwen3-1.7B, Type=Handcrafted2026.0533.7818
AutoTTS (β=1.0)Base Model=Qwen3-0.6B, Type=Discovered2026.0524.6480.9
Parallel-ProbeBase Model=Qwen3-0.6B, Type=Handcrafted2026.0524.1716.2
SC @ 64Base Model=Qwen3-0.6B, Type=Handcrafted2026.0523.2914.2
ASCBase Model=Qwen3-0.6B, Type=Handcrafted2026.0523.2617.3
ESCBase Model=Qwen3-0.6B, Type=Handcrafted2026.0523.2896.4
AutoTTS (β=0.5)Base Model=Qwen3-0.6B, Type=Discovered2026.0521.8254