MedFlowBench (Radiology) - Longitudinal MRI Strict: leaderboard
Metric: Strict score (0-1, shown times 100) on LUMIERE longitudinal brain MRI (139 questions: RANO response category from baseline and follow-up studies): the RANO category and every scorable follow-up lesion-state field (non-measurable lesion, T2 progression, current and new target and non-target lesions) must be right; the model drives 3D Slicer through the MedOpenClaw runtime with viewer-native actions only (series selection, scrolling, windowing, prealigned fusion), inspecting the complete study within a 20-round budget and returning an answer with structured evidence (Track A); higher is better. Source: arxiv.org. Saturation forecast: Around 2032. 14 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.7 | 18 | #45 |
| 2 | Qwen 3.5 35B A3B | 18 | #250 |
| 3 | Claude Sonnet 4.6 | 17 | #85 |
| 4 | Qwen 3.5 4B | 11 | #470 |
| 5 | Qwen 3.5 27B | 9 | #183 |
| 6 | Gemini 3.1 Pro (Preview) | 8 | #54 |
| 7 | Qwen 3.5 9B | 8 | #363 |
| 8 | Gemini 3 Flash (Preview) | 7 | #78 |
| 9 | GPT-5.4 Mini | 5 | #202 |
| 10 | GPT-5.5 | 4 | #26 |
| 11 | MedGemma 1.5 4B | 2 | #775 |
| 12 | Gemma 3 27B (IT) | 1 | #509 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=medflowbench-radiology-longitudinal-mri-strict · How It Works · Data refreshed daily, snapshot 2026-10-11.