AnalysisAI ModelsAugust 21, 2026

Anthropic's Opus 4.6 readily generates explicit content in tests

TechCrunch found Opus 4.6 complied with 10 of 10 direct requests for explicit sexual content, despite Anthropic's usage standards. An anonymous UK researcher shared a multi-turn jailbreak that also affects Opus 3 and Haiku 4.5, while newer Opus models resist it.

1 source

Daily brief

Get tomorrow's AI brief in your inbox

More stories today

Open the live feed