MedBench v5: leaderboard
Existing medical AI benchmarks lack process visibility, atomic skill evaluation, and integrated hallucination detection. We introduce MedBench v5, a redesigned benchmark for clinical multimodal models (language, vision-language, and agent systems) that moves from static QA to dynamic, process-oriented evaluation. MedBench v5 features: (1) a dual-dimensional framework combining Clinical Cognitive R
Metric: CCR-Agent (self-reported). Source: benchmarklist.com. Saturation forecast: Around August 2027. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.7 | 96.66 | #49 |
| 2 | Kimi K2.6 | 96.49 | #104 |
| 3 | GPT-5.5 | 95.09 | #31 |
| 4 | GLM-5.1 | 94.89 | #97 |
| 5 | Grok 4.20 | 94.6 | #175 |
| 6 | Gemini 3.1 Pro (Preview) | 94.42 | #58 |
| 7 | DeepSeek V4 Pro | 93.12 | #102 |
| 8 | Seed 2.0 Pro | 92.44 | #118 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=medbench-v5 · How It Works · Data refreshed daily, snapshot 2026-10-20.