Anthropic Disordered Eating Single-Turn Risk Requests - Harmless Rate (Claude.ai): leaderboard
Metric: Harmless Response Rate (%). Source: www-cdn.anthropic.com. 6 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.8 | 99.7 |
| 2 | Claude Opus 5 | 99.67 |
| 3 | Claude Fable 5 | 99.64 |
| 4 | Claude Sonnet 5 | 99.55 |
| 5 | Claude Fable 5.1 | 99.43 |
| 6 | Claude Sonnet 4.6 | 98.63 |
Interactive version: theaggregate.ai/benchmark?slug=anthropic-disordered-eating-single-turn-risk-requests-harmless-rate-claude-ai · How It Works · Data refreshed daily, snapshot 2026-09-19.