Claude Mythos Preview
Preview language model
A Claude-branded preview language model release associated with Anthropic, identified from the provided target name.
Pricing
Research preview model; no separate cached-input price was explicitly stated in the cited sources. Pricing is described as available to Project Glasswing participants and via Anthropic’s platform partners.
Anthropic’s official Project Glasswing page states that Claude Mythos Preview is available at $25 per million input tokens and $125 per million output tokens. Anthropic’s pricing docs also list Claude Mythos Preview under long-context pricing at standard pricing.
Model Intelligence
Recent stories
Multiple summaries of the UK AISI report say GPT-5.5 roughly matches Claude Mythos Preview on long-horizon cyber tasks, including 2 of 10 end-to-end TLO completions. That matters because the model is broadly usable today, shifting cyber-workflow choices toward availability and mitigations rather than gated access alone.
Posts summarizing WSJ reporting say Anthropic’s push to widen Mythos preview access by about 70 organizations was opposed over national-security and compute-capacity concerns. The change matters because access to Anthropic’s top cyber model may stay tightly rationed for defenders, vendors, and evaluators.
Anthropic's Mythos system card says the model completed the AI Security Institute's 32-step corporate attack range in about 20 human hours. The benchmark matters as a cyber capability signal, but the range is easier than a real defended enterprise network.
UK regulators put Claude Mythos on formal briefing agendas while US officials also pushed banks to evaluate it. Watch the independent critiques of Anthropic's exploit method, low-level access behavior, and small-model comparisons before treating the release as production-ready.