Anthropic: Tests Found Claude Fable 5 Posed No Unique Cybersecurity Risk

Anthropic said internal testing found Claude Fable 5 posed no unique cybersecurity danger as Claude Mythos 5 is set to return globally on July 2. According to BeInCrypto, the admission comes with Fable 5’s global relaunch after an 18-day suspension triggered by US export controls on June 12. Anthropic said tests showed Claude Opus 4.8, GPT-5.5, and Kimi K2.7 could identify the same vulnerabilities cited in an Amazon report and reproduce a single exploit demonstration. Anthropic said it redeployed Fable 5 with new classifiers that block the bypass in over 99% of cases and reroute blocked requests to Opus 4.8.