Anthropic releases Claude Fable 5.1 and Mythos 5.1, one model with two safety levels

On September 1, 2026, Anthropic released Claude Fable 5.1 and Claude Mythos 5.1. The company says both are the same underlying model with different safeguards, which are the rules and filters that limit what an AI will help with. Fable 5.1 is generally available, including through Anthropic's API (the connection developers use to build apps) and cloud platforms. Mythos 5.1 is limited to vetted professionals in cybersecurity and life sciences through special access programs. Anthropic says Fable 5.1 should cost about 25% less than Fable 5 for typical work, and it cut the price of cache reads, when the model reuses text it has already processed, by 75%. The update also tries to reduce false alarms. Anthropic reports that its cyber safeguards step in about 60% less often per session, and its biology safeguards trigger 85% less often on harmless basic questions. TechCrunch noted that the system card, a safety report published with the model, says Mythos 5.1 did slightly worse than earlier versions on an overall measure of unwanted behavior. Why it matters: AI companies must balance blocking dangerous requests against frustrating people with ordinary questions, and this release shows that trade-off in action. Source: Anthropic — https://www.anthropic.com/claude-fable-and-mythos-5-1 Source: TechCrunch — https://techcrunch.com/2026/09/01/anthropics-new-fable-release-is-cheaper-less-restrictive/