EHRNote-ChatQA: leaderboard
EHRNote-ChatQA (2026) poses multi-turn 5-way multiple-choice questions over the full sequence of MIMIC-IV discharge summaries of one patient, pairing each content question with an evidence-grounding question about which notes and sections support the answer. This board is paired accuracy in percent: both the content answer and its evidence answer must be correct, averaged over all turns; higher is better.
Metric: QA-level paired accuracy (%): the content answer and its paired evidence-grounding answer both correct, micro-averaged over all turns, on 967 multi-turn samples, each over the full set of MIMIC-IV discharge summaries of one patient (1 to 5 notes), 8,036 content and 8,036 paired evidence-grounding 5-way multiple-choice questions (chance 20), chat history kept across turns, greedy decoding where settable; higher is better. Source: arxiv.org. Saturation forecast: Around February 2027. 22 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.4 | 87.95 | #123 |
| 2 | Gemini 3 Flash (Preview) | 87.08 | #144 |
| 3 | Qwen 3 Next 80B A3B Instruct | 79.22 | #766 |
| 4 | Llama 4 Scout Instruct | 71.34 | #1543 |
| 5 | Qwen 3 30B A3B 2507 Instruct | 70.78 | #1089 |
| 6 | GPT-5.4 Mini | 68.96 | #389 |
| 7 | DeepSeek R1 Distill Llama 70B | 62.22 | #1637 |
| 8 | Qwen 3 4B 2507 Instruct | 55.99 | #2163 |
| 9 | DeepSeek R1 Distill Qwen 32B | 53.29 | #1788 |
| 10 | DeepSeek R1 Distill Qwen 14B | 47.5 | #2582 |
| 11 | Ministral-3-3B-Instruct-2512 | 44.3 | #2579 |
| 12 | Phi-4 Mini Instruct | 36.78 | #6419 |
| 13 | Phi-3.5-mini-instruct | 32.53 | #5733 |
| 14 | Ministral-3-8B-Instruct-2512 | 31.89 | #1858 |
| 15 | DeepSeek R1 Distill Llama 8B | 28.47 | #6754 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=ehrnote-chatqa · How It Works · Data refreshed daily, snapshot 2026-11-13.