Code Generation on HumanEval (pass@1, pass@2)
86pass@1EXG-SE
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| EXG-SEModel=Qwen3-Coder-Flash2026.05 | 86 | 95.1 | |
| EXG-SE-RevModel=Qwen3-Coder-Flash2026.05 | 83.5 | 92.1 | |
| EXG-ReflexionModel=Qwen3-Coder-Flash2026.05 | 81.7 | 87.2 | |
| EXGModel=Qwen3-Coder-Flash2026.05 | 80.5 | 83.5 | |
| SE-Agent-RevModel=Qwen3-Coder-Flash2026.05 | 64 | 94.5 | |
| SE-AgentModel=Qwen3-Coder-Flash2026.05 | 59.8 | 93.9 | |
| EXG-SE-RevModel=Qwen3-8B2026.05 | 59.1 | 79.9 | |
| ReflexionModel=Qwen3-Coder-Flash2026.05 | 58.5 | 78 | |
| EXG-ReflexionModel=Qwen3-1.7B2026.05 | 57.3 | 69.5 | |
| EXG-SEModel=Qwen3-8B2026.05 | 57.3 | 83.5 | |
| EXGModel=Qwen3-1.7B2026.05 | 53.7 | 58.5 | |
| EXG-SE-RevModel=Qwen3-1.7B2026.05 | 53.7 | 59.8 | |
| EXG-SEModel=Qwen3-1.7B2026.05 | 51.2 | 61 | |
| EXG-ReflexionModel=Qwen3-8B2026.05 | 50.6 | 72 | |
| EXGModel=Qwen3-8B2026.05 | 41.5 | 48.8 | |
| ReflexionModel=Qwen3-1.7B2026.05 | 20.7 | 54.3 | |
| SE-Agent-RevModel=Qwen3-1.7B2026.05 | 20.7 | 56.1 | |
| ReflexionModel=Qwen3-8B2026.05 | 20.1 | 61 | |
| SE-AgentModel=Qwen3-1.7B2026.05 | 19.5 | 52.4 | |
| SE-Agent-RevModel=Qwen3-8B2026.05 | 18.9 | 82.3 | |
| SE-AgentModel=Qwen3-8B2026.05 | 17.7 | 72 |