TriageBench: leaderboard
Clinical chart triage requiring the correct care decision from supplied patient information. Reports decision accuracy and retains reasoning effort and dated model identifiers. The main board contains 100-case evaluations; partial runs appear on boards labelled with their smaller case count. This board reports TriageBench using Decision Accuracy (%).
Metric: Decision Accuracy (%). Source: labs.logarith.ms. Saturation forecast: Not forecast. 29 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.8 | 45 | #39 |
| 2 | Claude Opus 5 | 43 | #16 |
| 3 | Gemini 3.1 Pro (Preview) | 42 | #64 |
| 4 | GPT-5.6 Sol | 42 | #22 |
| 5 | Qwen 3.8 Max | 41 | #34 |
| 6 | Qwen 3.7 Plus | 41 | #98 |
| 7 | Claude Opus 4.7 | 40 | #51 |
| 8 | Claude Sonnet 5 | 40 | #65 |
| 9 | GPT-5.6 Sol (Max) | 40 | #22 (GPT-5.6 Sol) |
| 10 | Muse Spark 1.1 (xHigh) | 40 | #55 (Muse Spark 1.1) |
| 11 | GPT-5.4 | 39 | #83 |
| 12 | GPT-5.5 | 39 | #33 |
| 13 | GPT-5.6 Luna | 39 | #79 |
| 14 | GPT-5.6 Terra (Max) | 39 | #40 (GPT-5.6 Terra) |
| 15 | Mistral Large 3 | 38 | #410 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=triagebench · How It Works · Data refreshed daily, snapshot 2026-11-05.