X-PCR - Attending-Level Questions: 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 on questions rated at attending difficulty, in percent; higher is better.

Metric: Accuracy (%) on the questions ophthalmologists rated at attending difficulty, 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: Around February 2027. 21 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1GPT-576.19#178
2Qwen 3 VL 30B A3B70.84#784
3Gemini 2.5 Pro70.68#276
4InternVL3-32B70.37
5Qwen 2.5 VL 72B70.11#781
6GPT-5 Nano69.28#906
7GLM-4.5V69.19#754
8Qwen 2.5 VL 7B68.43#1862
9InternVL3-8B67.3#1782
10Gemini 2.5 Flash65.48#463
11Lingshu-7B65.15#1635
12MedGemma-27B-IT64.63#1350
13Qwen 3 VL 8B61.49#917
14Claude Haiku 4.559.89#583
15HuaTuoGPT-Vision-34B59.31#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-attending-level-questions · How It Works · Data refreshed daily, snapshot 2026-11-21.