Anthropic adds security audit and pricing details for Claude Fable 5.2

Anthropic says Fable 5.2 and Mythos 5.2 improve on earlier versions in coding, reasoning and vision.[1] It describes typical uses as agent sessions lasting hours, multistep deep research, and analysis that produces a finished document, spreadsheet or deck.[2] The company says Fable 5.2 solved two open theoretical-physics problems involving black holes.[1] In another example, it says the model coordinated more than 1,000 subagents for eight days to write, test and review code.[1] These are company-reported examples, not a guarantee of results on a particular team’s work.

Fable 5.2 is generally available, while Mythos 5.2 is limited to vetted organizations through Anthropic’s Life Sciences and Cyber Verification Programs.[3] Anthropic lists Fable 5.2 at $8 per million input tokens and $40 per million output tokens, and says the models cost 20% less than the 5.1 generation.[3] Its optional fast mode is priced at $16 per million input tokens and $80 per million output tokens, with token generation up to 2.5 times faster, according to the company.[1] Platform documentation lists a one-million-token context window and a maximum output of 128,000 tokens.[2]

Anthropic says the new models generate outputs up to 30% faster than Fable 5.1 and Mythos 5.1.[1] At Xhigh effort, the company reports a 49.1% score for Fable 5.2, the highest score among released models it tested.[1] At High effort, it says Fable 5.2 scores about the same as Opus 5.5’s best score for roughly half the cost per task, but recommends starting with Opus 5.5 for most coding and knowledge work.[1] Teams should assess the specific task and operating requirements rather than treating benchmark results as a universal model choice.

Anthropic says Fable 5.2 ran a 10–12-hour security audit, dispatched about 170 reviewer agents, and returned a write-up, fix pull requests and a policy proposal.[1] Separately, the company says one security firm found that one in six confirmed bugs identified by Mythos 5.2 were missed by all the other frontier models it tested.[1] Anthropic says Fable’s cybersecurity and biology safeguards are as robust as Fable 5.1’s while causing fewer unnecessary blocks.[1] For organizations evaluating deployment, its documentation says Fable 5.2 requires 30-day data retention and is unavailable under zero data retention unless Anthropic expressly authorizes it.[2]