Medical Visual Question Answering on SLAKE n=350 (test)
86.3LLM Judge AccuracyDiffusionGemma
Evaluation results
| Method | Links | |
|---|---|---|
| DiffusionGemmaEvaluation Protocol=per-dataset finetuned, Backbone=diff2026.07 | 86.3 | |
| Gemma-4Evaluation Protocol=per-dataset finetuned, Backbone=AR2026.07 | 81.7 | |
| Gemini-3.5-FlashEvaluation Protocol=zero-shot, Model Category=frontier VLMs2026.07 | 75.1 | |
| Claude-Sonnet-4.6Evaluation Protocol=zero-shot, Model Category=frontier VLMs2026.07 | 70.3 | |
| DiffusionGemmaEvaluation Protocol=base (zero-shot), Backbone=diff2026.07 | 70 | |
| Gemma-4Evaluation Protocol=base (zero-shot), Backbone=AR2026.07 | 67.4 | |
| GPT-4.1-miniEvaluation Protocol=zero-shot, Model Category=frontier VLMs2026.07 | 64.6 |