Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

First-error detection on ProcessBench

68.7Accuracy

Teacher

22.62834.58946.5558.511May 13, 2026
Updated 20d ago

Evaluation Results

MethodLinks
2026.05
68.7
2026.05
46.3
2026.05
43.8
43.2
2026.05
34.4
24.4