MedBench v5 - CCR-Multimodal: leaderboard
Metric: Macro-average score (0-100) over the 12 tasks of the MedBench v5 Clinical Cognitive Responsiveness multimodal track (lesion detection, image classification, report OCR, visual question answering, report generation, longitudinal imaging and multimodal clinical decision support), each task normalized to 0-100; higher is better. Source: arxiv.org. Saturation forecast: Around September 2027. 10 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Seed 2.0 Pro | 51.65 |
| 2 | Gemini 3.5 Flash | 50.68 |
| 3 | GPT-5.5 | 50.07 |
| 4 | Qwen 3.7 Plus | 49.24 |
| 5 | Claude Opus 4.7 | 47.98 |
| 6 | Kimi K2.6 | 46.4 |
| 7 | GLM-5.1 | 42.35 |
| 8 | DeepSeek V4 Pro | 41.18 |
| 9 | Grok 4.20 | 35.09 |
Interactive version: theaggregate.ai/benchmark?slug=medbench-v5-ccr-multimodal · How It Works · Data refreshed daily, snapshot 2026-09-29.