SignalFrontier releaseSG-0003
Anthropic released Claude Mythos 5 to vetted groups and a safeguarded twin, Fable 5, to everyone
Anthropic released Claude Mythos 5 to approved security firms, infrastructure operators, government partners and some life-science researchers. Claude Fable 5, the same model, is for general use. It hands requests about cyberattacks, biology or chemistry, or about copying the model, to a less capable Claude model instead of answering them itself.
Why it matters
Routing risky requests to a weaker model is a new kind of safeguard, and a public admission that the full model is too capable to hand out freely. How well that routing holds up, and whether agents built on Fable find ways around it, is worth watching.
Sources
Read the reporting
2 sources. Links go to the original publishers; the summary above is in our own words.
- primaryClaude Fable 5 and Claude Mythos 5Anthropic · June 9, 2026anthropic.com/news/claude-fable-5-mythos-5
- newsAnthropic rolls out Claude Mythos 5 AI model, along with the safer Fable 5 for the publicDecrypt · June 9, 2026decrypt.co/370556/anthropic-rolls-out-claude-mythos-5-ai-model-along-with-the-s…
Signals
More from the news desk
Cite and share
Use this signal
Citation
Paperclip Index. “Anthropic released Claude Mythos 5 to vetted groups and a safeguarded twin, Fable 5, to everyone.” Signal SG-0003. Reported June 9, 2026; updated Oct. 7, 2026. https://paperclipindex.com/signal/SG-0003