Open-Web Reasoning on GAIA Open-web via SERPER (test)
46.8Pass@4Qwen3-8B
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen3-8BSampling Method=DivInit (DI)2026.06 | 46.8 | |
| Gemma3-12BSampling Method=DivInit (DI)2026.06 | 45.2 | |
| Qwen3-4BSampling Method=DivInit (DI)2026.06 | 44.9 | |
| Qwen3-8BSampling Method=Standard parallel sampling (S)2026.06 | 41.6 | |
| Qwen3-4BSampling Method=Standard parallel sampling (S)2026.06 | 38.7 | |
| Gemma3-12BSampling Method=Standard parallel sampling (S)2026.06 | 38 | |
| Gemma3-12BSampling Method=DivInit (DI)2026.06 | 35 | |
| Gemma3-12BSampling Method=Standard parallel sampling (S)2026.06 | 34 | |
| Gemma3-12BSampling Method=DivInit (DI)2026.06 | 31.6 | |
| Qwen3-8BSampling Method=DivInit (DI)2026.06 | 30.2 | |
| Qwen3-4BSampling Method=DivInit (DI)2026.06 | 29 | |
| Qwen3-8BSampling Method=DivInit (DI)2026.06 | 28.2 | |
| Gemma3-12BSampling Method=Standard parallel sampling (S)2026.06 | 28.2 | |
| Qwen3-4BSampling Method=DivInit (DI)2026.06 | 27.8 | |
| Qwen3-8BSampling Method=Standard parallel sampling (S)2026.06 | 26 | |
| Qwen3-8BSampling Method=Standard parallel sampling (S)2026.06 | 25.2 | |
| Qwen3-4BSampling Method=Standard parallel sampling (S)2026.06 | 23.7 | |
| Qwen3-4BSampling Method=Standard parallel sampling (S)2026.06 | 22.7 |