Coding on LiveCodeBench (ACC, UT)
46.7AccuracyQwen3-8B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Qwen3-8BInference Mode=think2026.03 | 46.7 | — | |
| SSA-LLM-8BInference Mode=think2026.03 | 40.11 | — | |
| RAM2026.01 | 31.96 | 47.72 | |
| DARE+TA2026.01 | 31.95 | 46.69 | |
| RAM+2026.01 | 31.6 | 46.84 | |
| WUDI2026.01 | 30.04 | 44.48 | |
| Qwen3-8BInference Mode=no-think2026.03 | 23.08 | — | |
| SSA-LLM-8BInference Mode=no-think2026.03 | 19.78 | — |