BanglaMedVQA (English): leaderboard
BanglaMedVQA (Ahmed et al., 2026) asks short-answer medical visual questions about ChestX-ray14 radiographs and MedICaT figures (modality, organ, abnormality, condition, position); QA pairs were generated from image metadata by Gemini 2.5 Flash and translated to Bangla by Gemini 2.0 Flash. This board uses the English source questions with the zero-shot prompt and scores exact-match accuracy in percent, the mean of the two 1,000-question subsets; higher is better.
Metric: Exact-match accuracy (%; mean of the Chest X-Ray and MedICaT subsets, 1,000 questions each; English questions, zero-shot prompt; exact match with the reference answer). Source: arxiv.org. Saturation forecast: Estimated already saturated. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 2.5 VL 7B Instruct | 48.25 | #1842 |
| 2 | Gemini 2.5 Flash | 45 | #452 |
| 3 | GPT-4.1 Mini | 37 | #775 |
| 4 | Gemma 3 12B | 33 | #2013 |
| 5 | MedGemma-4B | 24.5 | #2460 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=banglamedvqa-english · How It Works · Data refreshed daily, snapshot 2026-11-12.