True/False Question Answering on PBSBench Slide-level 1.0 (test)
86AccuracyPBS-VL
Evaluation Results
| Method | Links | |
|---|---|---|
| PBS-VLFramework=Slide-level, Vision backbone=Dinobloom ViT-L/14, VLM backbone=Qwen 2.5 VL-7B2026.04 | 86 | |
| SlideChatModel type=Pathology-specific for WSIs2026.04 | 77 | |
| Claude-4.5Input=30 random patches2026.04 | 69 | |
| GPT-4oInput=30 random patches2026.04 | 57 | |
| Gemini-2.5-proInput=30 random patches2026.04 | 49 | |
| HistoGPTModel type=Pathology-specific for WSIs2026.04 | 20 |