SearchAgent - Natural Questions — leaderboard
Metric: Exact Match (%). Source: huggingface.co. 12 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | ReSeek-Qwen2.5-7b-Instruct | 46.9 |
| 2 | ZeroSearch-Qwen2.5-7b-Instruct | 43.6 |
| 3 | ReSeek-Qwen2.5-3b-Instruct | 41.5 |
| 4 | ZeroSearch-Qwen2.5-3b-Instruct | 41.4 |
| 5 | Search-R1-Qwen2.5-7b-Instruct | 39.3 |
| 6 | RAG-Qwen2.5-7b-Instruct | 34.9 |
| 7 | Search-R1-Qwen2.5-3b-Instruct | 34.1 |
| 8 | SFT-Qwen2.5-7b-Instruct | 31.8 |
| 9 | R1-Qwen2.5-7b-Instruct | 27 |
| 10 | Search-o1-Qwen2.5-7b-Instruct | 15.1 |
Interactive version: theaggregate.ai/benchmark?slug=searchagent-natural-questions · How the rankings work · Data refreshed daily, snapshot 2026-07-22.