# Medical Visual Question Answering on VQA-RAD n=350 (test) Source: [Medical Visual Question Answering on VQA-RAD n=350 (test)]() Collection index: [llms.txt]() Data updated: 2026-07-03T06:30:20.239Z Evaluation of model performance on the n=350 test split of the VQA-RAD dataset for Medical Visual Question Answering, specifically reporting LLM-judge accuracy. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Medical Visual Question Answering | | Dataset (including any stated split) | VQA-RAD n=350 (test) | | Metric | LLM-Judge Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 7 of 7 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [LLM-Judge Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | LLM-Judge Accuracy | Result date | Report paper | | --- | --- | --- | --- | | Gemini-3.5-Flash | 77.7 | Not recorded | 6a4741e4143ebdcacacc59f0 | | Claude-Sonnet-4.6 | 65.4 | Not recorded | 6a4741e4143ebdcacacc59f0 | | DiffusionGemma | 64.9 | Not recorded | 6a4741e4143ebdcacacc59f0 | | Gemma-4 | 64.9 | Not recorded | 6a4741e4143ebdcacacc59f0 | | DiffusionGemma | 61.4 | Not recorded | 6a4741e4143ebdcacacc59f0 | | GPT-4.1-mini | 57.1 | Not recorded | 6a4741e4143ebdcacacc59f0 | | Gemma-4 | 52.3 | Not recorded | 6a4741e4143ebdcacacc59f0 | ### Gemini-3.5-Flash LLM-Judge Accuracy: 77.7. Result ID: 6a4756df60b88bb33991d6e1. [Reporting paper]() Recorded evaluation configuration: {"Evaluation Protocol":"zero-shot","Model Category":"frontier VLMs"}. This record does not assert independent reproduction. ### Claude-Sonnet-4.6 LLM-Judge Accuracy: 65.4. Result ID: 6a4756df60b88bb33991d6e3. [Reporting paper]() Recorded evaluation configuration: {"Evaluation Protocol":"zero-shot","Model Category":"frontier VLMs"}. This record does not assert independent reproduction. ### DiffusionGemma LLM-Judge Accuracy: 64.9. Result ID: 6a4756df60b88bb33991d6df. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Evaluation Protocol":"per-dataset finetuned","Backbone":"diff"}. This record does not assert independent reproduction. ### Gemma-4 LLM-Judge Accuracy: 64.9. Result ID: 6a4756df60b88bb33991d6e0. [Reporting paper]() Recorded evaluation configuration: {"Evaluation Protocol":"per-dataset finetuned","Backbone":"AR"}. This record does not assert independent reproduction. ### DiffusionGemma LLM-Judge Accuracy: 61.4. Result ID: 6a4756df60b88bb33991d6dd. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Evaluation Protocol":"base (zero-shot)","Backbone":"diff"}. This record does not assert independent reproduction. ### GPT-4.1-mini LLM-Judge Accuracy: 57.1. Result ID: 6a4756df60b88bb33991d6e2. [Reporting paper]() Recorded evaluation configuration: {"Evaluation Protocol":"zero-shot","Model Category":"frontier VLMs"}. This record does not assert independent reproduction. ### Gemma-4 LLM-Judge Accuracy: 52.3. Result ID: 6a4756df60b88bb33991d6de. [Reporting paper]() Recorded evaluation configuration: {"Evaluation Protocol":"base (zero-shot)","Backbone":"AR"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.