Medical Information Response Audit (MIRA) - Completeness: leaderboard
Metric: Completeness loss (Q2), essential information missing against the per-seed checklist, LLM-judge rubric score (1-5), where higher is worse, averaged over the 4,320 MIRA prompts without mitigation: 60 medically reviewed low-risk seed questions crossed with language (Chinese, English), register and health-literacy signals, three prompt formats and three framings; lower is better. Source: arxiv.org. Saturation forecast: Around January 2027. 5 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | DeepSeek V4 Pro | 2.27 |
| 2 | Qwen 3.6 Plus | 2.34 |
| 3 | GPT-5.4 | 2.37 |
| 4 | Claude Sonnet 4.6 | 2.4 |
| 5 | Llama 3.3 70B Instruct | 3.03 |
Interactive version: theaggregate.ai/benchmark?slug=medical-information-response-audit-mira-completeness · How It Works · Data refreshed daily, snapshot 2026-10-07.