PBSBench - Slide Multiple Choice: leaderboard

PBSBench (2026) tests vision-language models on peripheral blood smear hematopathology, with questions about whole blood smear slides. Multiple-choice accuracy in percent; higher is better.

Metric: Accuracy (%, times 100) on PBSBench slide-level multiple-choice questions about whole peripheral-blood-smear slides (55 questions); proprietary models see 30 sampled patches per slide, whole-slide models their own input format; higher is better. Source: arxiv.org. Saturation forecast: Around February 2028. 5 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1GPT-4o75#754
2Gemini 2.5 Pro73#280
3SlideChat51
4Claude Sonnet 4.535#267
5HistoGPT24

No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=pbsbench-slide-multiple-choice · How It Works · Data refreshed daily, snapshot 2026-11-22.