Anthropic launched Claude Fable 5.2 and Mythos 5.2, charging $8/million input tokens and $40/million output tokens, down 20%.

Anthropic announced standard pricing of $8 per million input tokens and $40 per million output tokens for Claude Fable 5.2 and Mythos 5.2, versus $10 and $50 previously. Both also offer an optional fast mode at $16 per million input tokens and $80 per million output tokens; Fable 5.2 became available to all paid users, on Claude Platform and through major cloud platforms.

The company reports improved coding, reasoning and visual capabilities, with output speeds up to 30% faster than the preceding versions and performance above Opus 5.5 on almost all capabilities it evaluated. Fable 5.2 is described as handling harder software-engineering tasks, long-running projects and hardware-related work such as robot control, and as designed to coordinate more than 1,000 AI agents simultaneously; these are reported company capability claims.

Anthropic says Fable and Mythos share the same underlying model, with Fable retaining strict safeguards and Mythos reducing some safeguards for cybersecurity and life-sciences work under a vetted-access program. It reports expanded Mythos access for selected U.S. and overseas organizations, pre-release U.S. government review, testing by the UK AI Safety Institute, Frontier Design and METR, and new evaluation methods; its assessment that high-risk misalignment remains unlikely is a company judgment, not an independent guarantee.

In related hardware research, Anthropic describes work with circuit-board company Diode Computers that began more than a year earlier with tasks such as reading chip datasheets and identifying supporting components. It says earlier models struggled with reliable board layouts and compatible component designs, whereas Claude can now perform end-to-end electronic-hardware design and prototyping.