Term
Claude Mythos 5
Claude Mythos 5 (June 9, 2026) is Anthropic's most powerful model line above Opus — technically identical to Fable 5 but without its safety classifiers, distributed through Project Glasswing.
Claude Mythos 5 — explained in more detail
Claude Mythos 5 was introduced on June 9, 2026 alongside Claude Fable 5, superseding the earlier Mythos Preview. The Mythos line sits in a tier above the Opus range and is regarded as Anthropic’s most capable model family. According to Anthropic, Mythos 5 and Fable 5 are technically identical and differ only in their safeguards: where Fable 5 intercepts requests relating to cybersecurity, biology/chemistry or model distillation through classifiers and routes them to the weaker Opus, Mythos 5 answers them directly.
Industry estimates reported by the Financial Times put Mythos at roughly 8 trillion parameters versus about 5 trillion for Fable 5. In a cyber range evaluation by the UK AI Security Institute, Claude Mythos ranked first, ahead of Claude Opus 4.6.
Example / Practical use
Mythos 5 is not a general-purpose product model but is aimed specifically at defensive security work. Anthropic does not release it broadly but distributes it through the Project Glasswing program to cybersecurity defenders, critical-infrastructure operators and open-source maintainers. On June 2, 2026 Anthropic expanded access to around 150 organizations in more than 15 countries; the April 2026 launch cohort included over 40 companies, among them Microsoft, Apple, Google, AWS and the Linux Foundation. Access is arranged via Anthropic, AWS or Google Cloud account teams.
Distinction
Unlike Sonnet or Haiku, Mythos 5 cannot be booked publicly via the API. Its near-identical but safeguarded counterpart Fable 5, by contrast, is available on the public API and in Claude Code. Mythos 5 is therefore the unlocked variant of the same base, whose distribution Anthropic deliberately limits to vetted defender partners. Availability was also subject to export controls: following a U.S. government directive on June 12, 2026, access was temporarily suspended and restored in late June.
Discover more
AI Recap July 2026: Two Anthropic Models, a Government Review and the Open-Weight Pressure
July 2026 in AI: Sonnet 5 and Opus 5 from Anthropic, GPT-5.6 after a US review, Kimi K3 as open weights, and new GEO controls in Search Console.
GlossaryClaude Fable 5.1
Anthropic's flagship Claude Fable model, released on 1 September 2026. A frontier model for demanding reasoning, coding, research and long-running agents; 1M-token context, multimodal input, proprietary access via API and the Claude apps.
EncyclopediaComparing AI models — who builds what and how to choose
The major model families in 2026 at a glance. Who builds Claude, GPT, Gemini, Llama, Mistral, DeepSeek, Qwen — and which model to pick when.