Code Generation on LiveCode (test)
67.8AccuracyLLM-Debate
Evaluation results
| Method | Links | ||
|---|---|---|---|
| LLM-Debate2026.05 | 67.8 | 0.0167 | |
| AgentSlimming2026.05 | 61.7 | 0.0025 | |
| Cost-aware AFlow2026.05 | 57.9 | 0.0084 | |
| AFlow2026.05 | 55.3 | 0.0117 | |
| ADAS2026.05 | 52.1 | 0.009 | |
| Self-Refine2026.05 | 49.1 | 0.0023 | |
| Self-Consistency CoT2026.05 | 47.2 | 0.0079 | |
| CoT2026.05 | 46.4 | 0.0013 |