MedBench v5 - RAGAgent: leaderboard
A Medical Atomic Skills environment of MedBench v5 (2026): retrieval-augmented clinical question answering with query rewriting, re-ranking, conflict-aware evidence integration and source attribution, scored 0-100 by the environment metrics. Higher is better.
Metric: Environment score (0-100) on RAGAgent (retrieval-augmented clinical question answering with query rewriting, re-ranking, conflict-aware evidence integration and source attribution; mean average precision and agent-as-a-judge); higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 4 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.5 | 94.27 | #49 |
| 2 | Kimi K2.6 | 94.13 | #178 |
| 3 | Seed 2.0 Pro | 92.47 | #211 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=medbench-v5-ragagent · How It Works · Data refreshed daily, snapshot 2026-11-13.