Claude Fable 5 and Claude Mythos 5

Details

Claude Fable 5 is Anthropic's most capable generally available model to date — a frontier model matching the Mythos class but shipping with safeguards that restrict responses in cybersecurity-sensitive areas. Fable 5 tops benchmarks in software engineering, knowledge work, vision, and scientific research, with Anthropic noting its performance lead over previous models grows with task length and complexity. The model prices at $10/M input and $50/M output tokens, less than half the cost of Claude Mythos Preview.\n\nThe safeguards are tuned conservatively: queries touching sensitive topics (primarily cybersecurity) are redirected to Claude Opus 4.8, triggering in fewer than 5% of sessions on average. Anthropic acknowledged these will sometimes catch harmless requests and committed to reducing false positives as capabilities improve.\n\nClaude Mythos 5 is the same underlying model with safeguards lifted in specific cybersecurity areas, deployed initially through Project Glasswing in collaboration with the US government. Anthropic describes Mythos 5 as having 'the strongest cybersecurity capabilities of any model in the world,' with plans to expand through a broader trusted access program for vetted organizations.\n\nThe dual-launch structure — safe general use version plus a restricted high-capability version for vetted users — represents Anthropic's operational template for releasing frontier-class models. Note: shortly after this launch (June 12), the US government issued an export control directive suspending all access to both Fable 5 and Mythos 5 for foreign nationals (see related article).