Medical Information Response Audit (MIRA): leaderboard
Metric: Underinformative simplification (D3) (self-reported). Source: benchmarklist.com. 5 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Llama 3.3 70B | 2.95 |
| 2 | Claude Sonnet 4.6 | 2.32 |
| 3 | GPT-5.4 | 2.29 |
| 4 | Qwen 3.6 Plus | 2.25 |
| 5 | DeepSeek V4 Pro | 2.16 |
Interactive version: theaggregate.ai/benchmark?slug=medical-information-response-audit-mira · How It Works · Data refreshed daily, snapshot 2026-09-05.