PBSBench - Cell Multiple Choice: leaderboard
PBSBench (2026) tests vision-language models on peripheral blood smear hematopathology, with questions about single white blood cell crops. Multiple-choice accuracy in percent on in-domain cells; higher is better.
Metric: Accuracy (%, times 100) on PBSBench cell-level in-domain multiple-choice questions about single white-blood-cell crops from peripheral blood smears (1,088 questions); questions generated from GPT-4o morphology captions of hematopathologist-labeled cells; greedy decoding, model chat templates; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 16 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | InternVL3.5-8B | 91 | #1604 |
| 2 | GPT-4o | 90 | #751 |
| 3 | Gemini 2.5 Pro | 88 | #276 |
| 4 | Qwen 2 VL 7B Instruct | 81 | #2573 |
| 5 | PathGen-LLaVA | 78 | #1406 |
| 6 | Qwen 2.5 VL 7B Instruct | 76 | #1902 |
| 7 | Claude Sonnet 4.5 | 72 | #261 |
| 8 | llava-v1.6-mistral-7B | 52 | #3999 |
| 9 | DeepSeek-VL-7B | 51 | #7399 |
| 10 | LLaVA-1.6-Vicuna-7B | 46 | |
| 11 | Quilt-LLaVA | 36 | #4973 |
| 12 | Idefics3-8B-Llama3 | 35 | #5542 |
| 13 | LLaVA-Med | 29 | |
| 14 | Med-Flamingo | 27 | |
| 15 | PA-LLaVA | 24 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=pbsbench-cell-multiple-choice · How It Works · Data refreshed daily, snapshot 2026-11-21.