MedMosaic - Open-Ended Speech and Sound: leaderboard
MedMosaic (Rajgarhia et al., 2026): 1,215 open-ended questions on consultations with inserted clinical sounds. A GPT-5.1 judge scores correctness, relevance, completeness and clarity against a reference answer on 0-1 scales; the mean is scaled to 0-100 (higher is better).
Metric: Judge score (x 100; GPT-5.1 rates correctness, relevance, completeness and clarity against the reference answer on 0-1 scales, unweighted mean; 1,215 open-ended questions on consultations with inserted clinical sounds). Source: arxiv.org. Saturation forecast: Estimated already saturated. 13 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro | 73.5 | #262 |
| 2 | Gemini 2.5 Flash | 68.2 | #452 |
| 3 | Gemma 3n E4B (IT) | 51.1 | #2220 |
| 4 | Qwen2.5-Omni-7B | 47.6 | #1797 |
| 5 | Phi-4 Multimodal Instruct | 46.6 | #3914 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=medmosaic-open-ended-speech-and-sound · How It Works · Data refreshed daily, snapshot 2026-11-12.