MedStreamBench - Proactive Responsiveness: leaderboard
Metric: Responsiveness (%; timing of the first alert relative to the expected critical time in streaming proactive-monitoring rounds with responses no_alert, uncertain or alert, early, late and missed alerts penalized; 0-1 x100). Source: arxiv.org. Saturation forecast: Around December 2026. 9 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 2.5 Pro | 44.94 |
| 2 | InternVL3.5-8B | 8.99 |
| 3 | Qwen 2.5 VL 7B | 5.47 |
| 4 | Lingshu-7B | 5.14 |
Interactive version: theaggregate.ai/benchmark?slug=medstreambench-proactive-responsiveness · How It Works · Data refreshed daily, snapshot 2026-09-29.