Open-ended scientific discovery on BixBench open-ended discovery (evaluation set)
4.74ES ScoreOpenEvolve
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| OpenEvolve2026.06 | 4.74 | 1.31 | 192 | 1 | 0.5 | |
| Raw agent2026.06 | 4.27 | 2.51 | 212 | 29 | 13.7 | |
| SAGA2026.06 | 3.96 | 2.86 | 109 | 37 | 33.9 | |
| StatefulDiscovery2026.06 | 3.75 | 3.14 | 136 | 33 | 24.3 | |
| AutoDiscovery2026.06 | 3.35 | 2.47 | 198 | 18 | 9.1 |