SearchAgent - FictionalHotpotQA: leaderboard
SearchAgent FictionalHotpotQA slice measures exact match on counterfactual multi-hop QA.
Metric: Exact Match (%). Source: huggingface.co. 12 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | ReSeek-Qwen2.5-7b-Instruct | 6.1 | |
| 2 | ReSeek-Qwen2.5-3b-Instruct | 5.9 | |
| 3 | Search-R1-Qwen2.5-3b-Instruct | 3.7 | |
| 4 | Search-R1-Qwen2.5-7b-Instruct | 3.4 | |
| 5 | ZeroSearch-Qwen2.5-7b-Instruct | 3.1 | |
| 6 | ZeroSearch-Qwen2.5-3b-Instruct | 3 | |
| 7 | Search-o1-Qwen2.5-7b-Instruct | 2 | |
| 8 | RAG-Qwen2.5-7b-Instruct | 1.2 | |
| 9 | R1-Qwen2.5-7b-Instruct | 0.3 | |
| 10 | SFT-Qwen2.5-7b-Instruct | 0.3 |
Interactive version: theaggregate.ai/benchmark?slug=searchagent-fictionalhotpotqa · How It Works · Data refreshed daily, snapshot 2026-10-20.