Multiple Choice Question Answering on PBSBench Slide-level 1.0 (test)
85AccuracyPBS-VL
Evaluation results
| Method | Links | |
|---|---|---|
| PBS-VLFramework=Slide-level, Vision backbone=Dinobloom ViT-L/14, VLM backbone=Qwen 2.5 VL-7B2026.04 | 85 | |
| GPT-4oInput=30 random patches2026.04 | 75 | |
| Gemini-2.5-proInput=30 random patches2026.04 | 73 | |
| SlideChatModel type=Pathology-specific for WSIs2026.04 | 51 | |
| Claude-4.5Input=30 random patches2026.04 | 35 | |
| HistoGPTModel type=Pathology-specific for WSIs2026.04 | 24 |