Planning on Natural Plan (in-distribution)
91.7Success RateGemini-3-Flash
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Gemini-3-FlashOrchestrator=prompted2026.06 | 91.7 | 3.2 | 5.4 | 0.08 | |
| HALOOrchestrator=ours2026.06 | 88.8 | 2.8 | 3.2 | 0.004 | |
| GPT-5-miniOrchestrator=prompted2026.06 | 51.5 | 4 | 6.8 | 0.22 |