When Symptoms Are Not Enough: leaderboard
As demand for mental health care outpaces clinician-delivered assessment, scalable screening tools are increasingly needed. Large language models (LLMs) may identify psychiatric risk from patient narratives, but their reliability across diagnoses, demographic subgroups, and evidence-use patterns remains uncertain. We introduce a SCID-anchored benchmark of 555 semi-structured experiential interview
Metric: Matthews correlation coefficient (self-reported). Source: benchmarklist.com. Saturation forecast: Around December 2026. 5 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4.1 Mini | 38 | #355 |
| 2 | GPT-5 Mini | 34 | #186 |
| 3 | GPT-4.1 | 30 | #248 |
| 4 | Gemini 2.5 Pro | 28 | #152 |
| 5 | Llama 4 Scout | 16 | #656 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=when-symptoms-are-not-enough · How It Works · Data refreshed daily, snapshot 2026-10-20.