SAFE-Teen Health Advice (under-18 prompt)
Safety
Pooled criterion failure rate for health advice and clinical boundaries, with the under-18 system prompt. The source suite has 34 scenarios in this group; branching conversations and omitted unjudged responses make the scored criterion denominator model dependent.
Top models (lower is better)
| Model | Score |
|---|---|
| Muse Spark 1.3 | 0.8 |
| GPT-6.1 Sol | 2.0 |
| GPT-5.6 Luna | 6.1 |
| Grok 4.5 | 6.4 |
| Grok 4.7 | 10.6 |
| Claude Sonnet 5.5 | 14.7 |
| Claude Opus 5.5 | 15.1 |
| Gemini 4 Argon | 29.6 |
| Gemini 3.6 Flash | 31.3 |
| Gemini 3.8 Flash | 36.1 |