Synthetic Hospital - Imaging Indication: leaderboard
Metric: Concept F1 (%; inferring the clinical question behind a vague imaging order, scored on graph-linked diagnosis and finding concepts, few-shot; public split of 200 synthetic longitudinal patients (1,268 in all, built from USMLE-style questions with ontology-grounded labels); single-turn, one prompting strategy locked per task for every model; graph-derived deterministic scoring). Source: arxiv.org. Saturation forecast: Around 2031. 10 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.3 | 51.8 |
| 2 | Kimi K2.5 (Thinking) | 51.7 |
| 3 | GLM-5 | 50.5 |
| 4 | Qwen 3.5 397B A17B | 49.7 |
| 5 | DeepSeek V3.2 | 49.2 |
| 6 | Claude Opus 4.6 | 47.3 |
| 7 | Gemma 3 27B | 41.8 |
| 8 | Llama 4 Scout | 40.6 |
Interactive version: theaggregate.ai/benchmark?slug=synthetic-hospital-imaging-indication · How It Works · Data refreshed daily, snapshot 2026-09-26.