Self-Verification on OlympiadBench
81.9Pass@1FABSVer-1.5B
Evaluation Results
| Method | Links | |
|---|---|---|
| FABSVer-1.5BBackbone=Qwen2.5-1.5B, Training=FABSVer2026.05 | 81.9 | |
| FABSVer-3BBackbone=Qwen2.5-3B, Training=FABSVer2026.05 | 77.8 | |
| RISE-1.5BBackbone=Qwen2.5-1.5B, Training=RISE2026.05 | 74.9 | |
| RISE-3BBackbone=Qwen2.5-3B, Training=RISE2026.05 | 73.6 | |
| FABSVer-7BBackbone=Qwen2.5-7B, Training=FABSVer2026.05 | 70.1 | |
| RISE-7BBackbone=Qwen2.5-7B, Training=RISE2026.05 | 65.9 | |
| GPT-4oModel=GPT-4o2026.05 | 54.4 | |
| Qwen2.5-1.5B BaseBackbone=Qwen2.5-1.5B, Training=Base2026.05 | 21.1 |