Agentic Search on Agentic-Wide WebWalkerQA (test)
62.5AccuracyA-MAPREDUCE
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| A-MAPREDUCEBackbone=GPT-5-mini2026.02 | 62.5 | 297.48 | 0.08 | |
| Flash-SearcherBackbone=GPT-5-mini2026.02 | 59.17 | 326.42 | 0.07 | |
| OWLBackbone=GPT-5-mini2026.02 | 52.5 | 652.08 | 0.11 | |
| SmolAgentsBackbone=GPT-5-mini2026.02 | 50 | 836.29 | 0.21 |