Code Generation on Mercury (test)
90.3Pass Rate (Easy)GCSL-bey
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| GCSL-bey2026.05 | 90.3 | 79.3 | 56 | 76.4 | 94.1 | 59 | 44.5 | 67.6 | |
| GCSL-NL2026.05 | 85.7 | 73.9 | 57.9 | 73.1 | 94.7 | 54.3 | 46.5 | 66.2 | |
| GCSL-bey-NL2026.05 | 84.5 | 82.8 | 64 | 77.7 | 95.6 | 65 | 64 | 71.4 | |
| GCSL2026.05 | 83.9 | 72.4 | 56 | 71.8 | 94.3 | 50 | 45.9 | 65 | |
| SFT2026.05 | 83.8 | 82.7 | 47.5 | 72.3 | 77.3 | 57.5 | 46.5 | 56.3 | |
| AlphaDPO2026.05 | 79.8 | 73.2 | 49.3 | 71 | 71.2 | 49.5 | 41.4 | 59.2 | |
| SimPO2026.05 | 78.1 | 70.5 | 42.6 | 69.8 | 66.8 | 48.1 | 38.7 | 53.4 | |
| DPO2026.05 | 77.8 | 68.7 | 41.3 | 67.5 | 64 | 45 | 37.4 | 49.8 |