MI-CXR - Trajectory Summary (Multiple Abnormalities): leaderboard

Trajectory Summary (Multiple Abnormalities) 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: compare the courses of several abnormalities across all visits. Accuracy in percent (chance 20); higher is better.

Metric: Global Trajectory Summarization comparing several abnormalities across all visits, 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 November 2027. 14 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1InternVL3.5-38B51.5#1021
2Gemini 3 Pro (Preview)40.7#119
3GPT-5.239#185
4InternVL3.5-8B37.1#1604
5Qwen 3 VL 32B32.5#622
6Claude Sonnet 4.529.2#261
7Lingshu-32B29#1113
8InternVL3.5-14B27.6#1504
9Qwen 3 VL 8B23.6#917
10MedGemma-27B-IT21.4#1350
11Lingshu-7B19.4#1635
12MI-CXR DeepSeek-VL-16B (checkpoint unspecified)18.7
13MedGemma-4B18.3#2557
14IDEFICS2-8B17.8#5013

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-trajectory-summary-multiple-abnormalities · How It Works · Data refreshed daily, snapshot 2026-11-21.