SurgCoT - Cue-Action Alignment: leaderboard

Aligning pre-action cues with micro-actions to pinpoint when and where an action begins in surgical videos of SurgCoT (2026), multiple-choice questions over 2,841 videos from seven specialties, answered zero-shot with only the video and the main question. Main-question accuracy in percent; higher is better.

Metric: Cue-action alignment main-question accuracy (%) on SurgCoT surgical-video multiple-choice questions (2,841 videos across 7 specialties and 35 procedures), baseline setting with only the video and the main question, zero-shot, temperature 0; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 10 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Claude Sonnet 4.583.66#261
2GPT-580.99#178
3Gemini 2.5 Pro77.97#276
4Qwen 3 VL 8B75.51#917
5MedGemma-27B-IT74.02#1350
6InternVL3.5-8B70.08#1604
7HuatuoGPT-Vision-7B68.28#2162
8LLaVA-Med-7B66.1#6296
9Lingshu-7B65.83#1635
10Qwen 2.5 VL 7B65.64#1862

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=surgcot-cue-action-alignment · How It Works · Data refreshed daily, snapshot 2026-11-21.