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

Iterative Debugging on OR-Debug-Bench

95.3RR@5

8B model (RLVR training)

85.83688.29390.7593.207Jan 28, 2026
Updated 5mo ago

Evaluation Results

MethodLinks
2026.01
95.362.42.25
2026.01
86.247.83.78