Anthropic announced Claude Fable 5 and Claude Mythos 5 on June 9, 2026. Fable 5 is a Mythos-class model made safe for general use and is described as the company’s most capable generally available model yet, achieving state-of-the-art results across benchmarks in software engineering, knowledge work, vision, scientific research, and other areas. Because of cybersecurity and misuse risks, Anthropic added conservative safeguards that route some sensitive queries to Claude Opus 4.8; these safeguards trigger in under 5% of sessions on average, though they may sometimes block harmless requests.
Claude Mythos 5 is the same base model, but with certain safeguards removed for a small group of cyber defenders and infrastructure providers. It will first be deployed through Project Glasswing with the U.S. government as an upgrade to Claude Mythos Preview, and Anthropic says it has the strongest cybersecurity capabilities of any model in the world. Broader trusted access is planned later.
The models are also positioned as useful for science and engineering. Early reports say Fable 5 can compress months of software work into days, perform complex code migrations, and handle long-horizon coding tasks more autonomously. It also shows strong results in financial reasoning, vision tasks, memory, and long-context work. Mythos 5 has helped with protein design, generated novel molecular biology hypotheses, and supported genomics research. Anthropic says automated alignment testing found low levels of misaligned behavior, similar to Opus 4.8. Both models are priced at $10 per million input tokens and $50 per million output tokens, less than half the price of Claude Mythos Preview.
Article timeline
Build this article’s timeline
Extract dated events, key actors, and open questions using the same timeline analysis available in Signal.
Comments
Login to view and post comments.