SearchAgent - MuSiQue: leaderboard
SearchAgent MuSiQue slice measures exact match on compositional multi-hop questions.
Metric: Exact Match (%). Source: huggingface.co. 12 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | ReSeek-Qwen2.5-7b-Instruct | 18.5 | |
| 2 | ZeroSearch-Qwen2.5-7b-Instruct | 18.4 | |
| 3 | Search-R1-Qwen2.5-7b-Instruct | 14.6 | |
| 4 | ReSeek-Qwen2.5-3b-Instruct | 10.3 | |
| 5 | Search-R1-Qwen2.5-3b-Instruct | 10.3 | |
| 6 | ZeroSearch-Qwen2.5-3b-Instruct | 9.8 | |
| 7 | R1-Qwen2.5-7b-Instruct | 7.2 | |
| 8 | SFT-Qwen2.5-7b-Instruct | 6.6 | |
| 9 | RAG-Qwen2.5-7b-Instruct | 5.8 | |
| 10 | Search-o1-Qwen2.5-7b-Instruct | 5.8 |
Interactive version: theaggregate.ai/benchmark?slug=searchagent-musique · How It Works · Data refreshed daily, snapshot 2026-10-20.