LoMeVQA - Progress Report Generation: leaderboard
LoMeVQA (2026) poses longitudinal visual questions about two or more chest radiographs of the same patient taken at different times. This board scores progress reports written from an image sequence by F1-RadGraph, the overlap of clinical entities and relations with the reference, in percent; higher is better.
Metric: F1-RadGraph clinical efficacy (%; progress report written from an image sequence, clinical entities and relations matched against the reference; LoMeVQA-test, 2,500 expert-reviewed longitudinal chest radiograph samples over five tasks, zero-shot; an input longer than the model window prints N/A and is left out; higher is better). Source: arxiv.org. Saturation forecast: Around 2030. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Flash | 14.28 | #453 |
| 2 | GPT-5 | 11.77 | #167 |
| 3 | GPT-4o | 10.79 | #741 |
| 4 | MedGemma-4B | 10.73 | #2444 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=lomevqa-progress-report-generation · How It Works · Data refreshed daily, snapshot 2026-11-13.