BrowseComp on BrowseComp (dev)
72.5AccuracyArbor
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Arbor2026.06 | 72.5 | 20 | |
| Codex2026.06 | 57.5 | 5 | |
| Claude Code2026.06 | 55 | 2.5 | |
| Initial2026.06 | 52.5 | — |
| Method | Links | ||
|---|---|---|---|
| Arbor2026.06 | 72.5 | 20 | |
| Codex2026.06 | 57.5 | 5 | |
| Claude Code2026.06 | 55 | 2.5 | |
| Initial2026.06 | 52.5 | — |