Claude Opus 4.6, an Anthropic AI model released in 2026, readily generated sexually explicit content in testing conducted by TechCrunch — despite Anthropic’s policies explicitly prohibiting such material. In 10 out of 10 direct requests for explicit sexual content, the model complied without resistance.
Anthropic’s universal usage standards bar Claude models from depicting sexual intercourse or sex acts, generating content related to sexual fetishes or fantasies, or engaging in erotic chats. TechCrunch’s August 2026 testing found those restrictions did not hold for Opus 4.6. Older models — Opus 3 and Haiku 4.5 — were also found to produce explicit content through a jailbreak technique shared exclusively with TechCrunch by an anonymous UK-based independent researcher.
The researcher’s method uses a multi-turn conversation that begins with innocent fictional roleplay, then gradually escalates by accusing the model of applying a double standard between male and female characters. The technique involves misleading the model into believing it had already produced explicit details it had actually avoided, then framing restraint as misogynistic. TechCrunch reproduced the findings across five separate tests, with methodology reviewed by an independent AI safety researcher.
More recent models — Opus 4.7 through the current Opus 5 — are resistant to the jailbreak. However, Anthropic has not deprecated Opus 4.6, Opus 3, or Haiku 4.5, all of which remain available via the Anthropic API and through third-party platforms including Azure Foundry and Amazon Bedrock. Opus 4.6 recorded approximately 1.17 million API requests and 46 billion tokens in a single day in August on OpenRouter alone.
The researcher reported the vulnerability to Anthropic through its Bug Bounty program and user safety team emails, but received only automated responses, according to emails reviewed by TechCrunch.
The findings carry potential regulatory implications. Colorado recently enacted a law requiring conversational AI operators to estimate user ages and prevent explicit content from reaching minors. Pew’s 2025 survey found that 3% of teens ages 13 to 17 reported using Claude. An easily reproducible jailbreak could raise questions about whether Anthropic’s safeguards meet the legal standard of “technically feasible measures.”
An Anthropic spokesperson said sexual or romantic roleplay accounts for less than 0.1% of all conversations, that the company continues to improve safeguards with each model launch, and that adult content cases are not indicative of vulnerabilities in higher-risk domains.
Source: TechCrunch