DRNoise (Clean): leaderboard
The clean condition of DRNoise (2026): accuracy in percent on 100 deep-research answer-recovery tasks whose gold answer must be assembled from two indirect record chains, run in a search-only BrowseComp-Plus agent without the misleading document. Higher is better.
Metric: Clean-condition accuracy (%; 100 deep-research tasks whose unique gold answer must be recovered from two corroborating indirect record chains, run in a fixed BrowseComp-Plus search-only ReAct agent (top-5 dense retrieval, 512-token snippets, up to 100 iterations); mean of three runs, without the misleading document; higher is better). Source: arxiv.org. Saturation forecast: Estimated already saturated. 5 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3.5 Flash | 96.3 | #101 |
| 2 | DeepSeek V4 Flash | 88.7 | #258 |
| 3 | MiMo-V2.5 | 82 | #354 |
| 4 | GPT-5.4 | 81.7 | #123 |
| 5 | Qwen 3 32B | 13 | #960 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=drnoise-clean · How It Works · Data refreshed daily, snapshot 2026-11-13.