Coding on LiveCodeBench (accuracy)
29.1AccuracySynthesized-Trace
Evaluation Results
| Method | Links | |
|---|---|---|
| Synthesized-TraceSurrogate=R1, Student Model=Qwen2026.03 | 29.1 | |
| No finetuningSurrogate=–, Student Model=Qwen2026.03 | 28.9 | |
| Synthesized-TraceSurrogate=R1-Distill, Student Model=Qwen2026.03 | 26.2 | |
| Answer-onlySurrogate=–, Student Model=Qwen2026.03 | 25.4 | |
| Surrogate-Trace(R1-Distill)Surrogate=–, Student Model=Qwen2026.03 | 22.9 | |
| No finetuningSurrogate=–, Student Model=Llama2026.03 | 13.1 | |
| Answer-onlySurrogate=–, Student Model=Llama2026.03 | 9.6 | |
| Synthesized-TraceSurrogate=R1-Distill, Student Model=Llama2026.03 | 9.2 | |
| Surrogate-Trace(R1-Distill)Surrogate=–, Student Model=Llama2026.03 | 9.1 | |
| Synthesized-TraceSurrogate=R1, Student Model=Llama2026.03 | 8.6 |