Synthetic Hospital - Specialty Summarization: leaderboard

Metric: Specialty-relevance F1 (%; specialty-conditioned summary scored against graph-derived specialty-relevant findings, zero-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 May 2027. 10 models tracked.

Top models

#ModelScore
1Claude Opus 4.668
2GLM-561.4
3GPT-5.359.5
4Kimi K2.5 (Thinking)59
5Qwen 3.5 397B A17B54.8
6DeepSeek V3.252.3
7Gemma 3 27B41.8
8Llama 4 Scout36.5

Interactive version: theaggregate.ai/benchmark?slug=synthetic-hospital-specialty-summarization · How It Works · Data refreshed daily, snapshot 2026-09-26.