Anthropic launched Claude Fable 5.2 and Claude Mythos 5.2 on Monday, offering lower token prices and claimed speed improvements while retaining verification requirements for direct access to Mythos.
Although the releases carry different names, Anthropic says they share the same underlying model. Their safeguards and access arrangements distinguish the two offerings.
Fable 5.2 targets demanding coding, reasoning and extended AI-agent tasks. Anthropic makes it more broadly available, with safeguards that restrict certain cybersecurity and biology activities.
Organisations seeking Mythos 5.2 for cybersecurity or biomedical research must obtain approval through the company’s verification programmes.
Anthropic says both versions generate outputs up to 30% faster than their 5.1 predecessors. It also reports clearer communication with users, based on feedback from prerelease testing.
Standard application programming interface (API) pricing stands at US$8 per million input tokens and US$40 per million output tokens.
For Fable, those rates represent a 20% reduction from the previous version’s US$10 input and US$50 output prices. Input tokens cover material the model receives, while output tokens cover the text it generates. An optional fast mode costs more.
Read More: Anthropic IPO to Grant Founders 50.1% Voting Power
Anthropic prices that mode at US$16 per million input tokens and US$80 per million output tokens, saying it can increase token-generation speed by up to 2.5 times.
Its developer documentation lists a one-million-token context window and a maximum output of 128,000 tokens for the new versions.
The company positions Fable and Mythos for more demanding, open-ended technical work. For most coding and knowledge tasks, it recommends starting with Claude Opus 5.5 before moving to another model, depending on the workload.
The launch follows Anthropic’s October release of Claude Haiku 5.5 as it expands its product range and prepares for a proposed initial public offering.
