Code Generation on LiveCodeBench Post-March 2025 V6 (Pass@1)
94.4Pass@1 (Easy)CodeSIM
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| CodeSIMModel=GPT-5-Mini2026.03 | 94.4 | 80 | 45.9 | 67.5 | |
| BACEModel=GPT-5-Mini2026.03 | 94.4 | 84 | 51.4 | 71.3 | |
| AgentCoderModel=GPT-5-Mini2026.03 | 88.9 | 52 | 35.1 | 52.5 | |
| DirectModel=GPT-5-Mini2026.03 | 83.3 | 60 | 27 | 50 | |
| MapCoderModel=GPT-5-Mini2026.03 | 83.3 | 80 | 35.1 | 60 | |
| CodeSIMModel=Qwen2.5-Coder:7b2026.03 | 72.2 | 16 | 8.1 | 25 | |
| BACEModel=Qwen2.5-Coder:7b2026.03 | 72.2 | 28 | 10.8 | 30 | |
| AgentCoderModel=Qwen2.5-Coder:7b2026.03 | 38.9 | 8 | 2.7 | 12.5 | |
| DirectModel=Qwen2.5-Coder:7b2026.03 | 27.8 | 4 | 0 | 7.5 |