Our new X account is live! Follow @wizwand_team for updates

Process Verification on ProcessBench Without Standard Answers

71.9Precise Accuracy

OPV-32B

Updated 4d ago

Evaluation Results

Method	Links
OPV-32B 2025.12		71.9	71.9	74.7	74.7	78.2	79.1	83.1	83.1
OPV-Stage2 2025.12		64.4	64.4	70.3	70.3	70.1	73.8	77.6	79
R1-Distill-Qwen-32B 2025.12		61.7	61.7	71.1	71.1	65	72.9	69.4	75.5
Qwen3-32B 2025.12		61.2	61.2	68.7	68.7	67	72.1	74.2	76.7
Qwen3-Max-Preview 2025.12		61	61	67.3	67.3	67	70.8	75.2	76.3
OPV-Stage1 2025.12		60.9	60.9	70.3	70.3	64.8	72.4	69.8	75.4
DeepSeek-V3-0324 2025.12		60.8	60.8	67.8	67.8	66	70.8	72.8	75.2
QwQ-32B 2025.12		58.2	58.2	66.7	66.7	64.2	70.1	70.9	74.2
gpt-oss-120b (high) 2025.12		57.9	57.9	64.1	64.1	66.8	69.3	75.6	75.4
DeepSeek-R1-0528 2025.12		56.9	56.9	64.7	64.7	63.2	68.2	71.7	73.6
Qwen2.5-Math-PRM-72B 2025.12		55.1	55.1	66	66	58.2	67.6	63.4	70.4
gpt-oss-120b (high) 2025.12		0.847	0.847	0.803	0.803	0.904	0.866	0.962	0.942
Qwen3-Max-Preview 2025.12		0.844	0.844	0.799	0.799	0.898	0.859	0.952	0.929
DeepSeek-R1-0528 2025.12		0.832	0.832	0.793	0.793	0.893	0.858	0.96	0.941
OPV-32B 2025.12		0.809	0.809	0.768	0.768	0.881	0.841	0.958	0.938
Qwen2.5-Math-PRM-72B 2025.12		0.767	0.767	0.732	0.732	0.839	0.798	0.912	0.878
R1-Distill-Qwen-32B 2025.12		0.758	0.758	0.736	0.736	0.836	0.805	0.933	0.911
DeepSeek-V3-0324 2025.12		0.721	0.721	0.698	0.698	0.81	0.772	0.915	0.884