Code Reasoning on LiveCodeBench v5
75.4Avg@8 ScoreR1-Distill-Qwen-32B
Evaluation Results
| Method | Links | |
|---|---|---|
| R1-Distill-Qwen-32B# Trainable Params=32B2025.10 | 75.4 | |
| s1.1-32B# Trainable Params=32B2025.10 | 58 | |
| Target + THINKLOGIT# Trainable Params=02025.10 | 54.1 | |
| Target + THINKLOGIT-DPO# Trainable Params=78M2025.10 | 52.4 | |
| Qwen2.5-32B# Trainable Params=-2025.10 | 48.3 | |
| R1-Distill-Qwen-1.5B# Trainable Params=-2025.10 | 36.2 |