List your company

Step-wise Verification on ProcessBench OlympiadBench

61.3TNRCons@4
5.55620.02834.548.972May 20, 2026

Evaluation results

MethodLinks
Cons@4Backbone=Qwen3-8B2026.0561.377.968.7
VerifySteer-BiBackbone=Qwen3-8B2026.0560.176.767.4
VerifySteer-UniBackbone=Qwen3-8B2026.0559.67767.2
Qwen3-14BBackbone=Qwen3-14B2026.0558.974.365.7
Prompt Opt.Backbone=Qwen3-8B2026.0558.477.966.7
Basic PromptBackbone=Qwen3-8B2026.055476.163.2
Qwen2.5-72BBackbone=Qwen2.5-72B2026.0535.987.650.9
Llama-3.3-70BBackbone=Llama-3.3-70B2026.0532.893.248.6
VerifySteer-UniBackbone=Qwen3-1.7B2026.0515.656.624.4
VerifySteer-BiBackbone=Qwen3-1.7B2026.051558.423.8
Prompt Opt.Backbone=Qwen3-1.7B2026.0513.349.621
Cons@4Backbone=Qwen3-1.7B2026.051158.518.5
Basic PromptBackbone=Qwen3-1.7B2026.057.778.814.1