MedProbeBench - Claim Recall: leaderboard
Expert-level medical guideline writing by LLMs with search tools on MedProbeBench (2026): each system writes a complete clinical guideline for 50 tumor entities from WHO classification volumes, compared with the expert guideline. Share of the expert atomic claims that the generated guideline covers, times 100; higher is better.
Metric: Task success rate (x100): share of the expert guideline atomic claims that the generated guideline covers (semantic matching, deduplicated), for LLMs with search tools on MedProbeBench: each system writes a complete clinical guideline for 50 WHO-classification tumor entities (one run per task), compared with the expert guideline; higher is better. Source: arxiv.org. Saturation forecast: Around March 2028. 8 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.2 | 35.3 | #185 |
| 2 | Claude Sonnet 4 | 34.2 | #410 |
| 3 | GPT-5 | 34 | #178 |
| 4 | Baichuan-M2-Plus | 33.8 | |
| 5 | Gemini 3 Flash (Preview) | 31.9 | #152 |
| 6 | GPT-4.1 | 30.1 | #505 |
| 7 | Grok 4 | 29.6 | #319 |
| 8 | Baichuan-M3-Plus | 28.3 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=medprobebench-claim-recall · How It Works · Data refreshed daily, snapshot 2026-11-21.