MedQ-Deg - Intensity Degradation: leaderboard
Metric: Accuracy (%) on MedQ-Deg items whose images carry intensity-jitter degradations, pooled over the L1 (mild) and L2 (severe) severities, multiple-choice medical VQA items merged from OmniMedVQA, GMAI-MMBench and MedXpertQA, each image corrupted by modality-specific degradations whose severity three radiologists calibrated (24,894 QA pairs over 7 modalities and 18 degradation types), single-letter answer, temperature 1.0, mean of 3 runs; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 40 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3 VL 30B A3B Instruct | 76.2 | #365 |
| 2 | Qwen 3 VL 235B A22B Instruct | 75.3 | #264 |
| 3 | GPT-5 | 72.2 | #91 |
| 4 | Gemini 2.5 Pro | 71.1 | #145 |
| 5 | GPT-5.1 | 70.9 | #131 |
| 6 | GPT-4.1 | 69.4 | #240 |
| 7 | GPT-4o | 68.8 | #333 |
| 8 | Gemini 2.5 Flash | 67.5 | #237 |
| 9 | GPT-4.1 Mini | 67.1 | #346 |
| 10 | MedGemma-4B | 66 | #731 |
| 11 | GLM-4.5V | 65.5 | #339 |
| 12 | GPT-4o Mini | 62.3 | #588 |
| 13 | Claude Sonnet 4.5 | 59.1 | #138 |
| 14 | Gemma 3 27B | 58.2 | #596 |
| 15 | Qwen 2.5 VL 72B Instruct | 57.6 | #364 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=medq-deg-intensity-degradation · How It Works · Data refreshed daily, snapshot 2026-10-11.