MI-CXR - Multiple Event Localization: leaderboard

Multiple Event Localization questions of MI-CXR (2026), five-option questions over five-visit chest X-ray timelines from MIMIC-CXR, answered zero-shot from all five images: identify the interval containing the requested occurrence of repeated appearance or resolution of one abnormality. Accuracy in percent (chance 20); higher is better.

Metric: Temporal Event Localization, multiple emergences or resolutions of one abnormality, five-option multiple-choice accuracy (%, times 100; chance 20) on MI-CXR, longitudinal reasoning over five-visit chest X-ray timelines from MIMIC-CXR (5,311 questions), zero-shot single-step prompting with all five images, temperature 0; higher is better. Source: arxiv.org. Saturation forecast: Around 2029. 14 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1GPT-5.237.1#185
2MedGemma-27B-IT35.1#1350
3Gemini 3 Pro (Preview)32.5#119
4IDEFICS2-8B30.8#5013
5InternVL3.5-38B30.6#1021
6InternVL3.5-8B29.5#1604
7Qwen 3 VL 8B29.5#917
8InternVL3.5-14B26.6#1504
9Lingshu-7B26#1635
10Lingshu-32B24.7#1113
11Qwen 3 VL 32B24.6#622
12Claude Sonnet 4.522.2#261
13MedGemma-4B19.6#2557
14MI-CXR DeepSeek-VL-16B (checkpoint unspecified)12.4

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

Interactive version: theaggregate.ai/benchmark?slug=mi-cxr-multiple-event-localization · How It Works · Data refreshed daily, snapshot 2026-11-21.