X-PCR - Image Quality Assessment: leaderboard

X-PCR (2026) tests multimodal models on ophthalmic diagnosis as a six-stage reasoning chain over six imaging modalities, with multiple-choice questions built from 51 public datasets and checked by ophthalmologists. Accuracy at the first stage, judging whether an image is good enough for diagnosis, in percent; higher is better.

Metric: Accuracy (%) at stage 1, judging whether image quality supports diagnosis, on X-PCR's ophthalmic multiple-choice VQA (26,415 images from 51 public datasets, six imaging modalities, 52 diseases), zero-shot, temperature 0; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 21 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1GPT-598.9#178
2Gemini 2.5 Pro94.58#276
3InternVL3-32B94.35
4Qwen 2.5 VL 72B93.13#781
5GPT-5 Nano91.3#906
6GLM-4.5V90.56#754
7MedGemma-27B-IT90.03#1350
8Gemini 2.5 Flash89.68#463
9Qwen 3 VL 8B89.4#917
10Qwen 3 VL 30B A3B89.25#784
11Qwen 2.5 VL 7B88.34#1862
12InternVL3-8B87.57#1782
13Lingshu-7B80.5#1635
14Claude Haiku 4.577.59#583
15HuaTuoGPT-Vision-34B75.94#1828

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

Interactive version: theaggregate.ai/benchmark?slug=x-pcr-image-quality-assessment · How It Works · Data refreshed daily, snapshot 2026-11-21.