Anthropic released Claude Fable 5.2 and Claude Mythos 5.2 on Monday, Nov. 9, an upgrade to the two models it describes as built for "the most ambitious, open-ended technical work." Fable 5.2 is available to everyone today, in Claude's apps and through Amazon Web Services, Google Cloud and Microsoft Azure. Mythos 5.2 is limited to vetted organizations.
The two are the same underlying model with different guardrails. Fable carries safety filters that block high-risk cybersecurity and biology requests and hand them to an older model, while Mythos drops those filters and goes only to cybersecurity and life sciences groups that pass Anthropic's verification programs.
What it costs
For developers, both models cost $8 per million input tokens and $40 per million output tokens, which Anthropic says is lower than before. An optional fast mode costs twice that. Anthropic says it generates text up to 2.5 times faster.
For people who pay for the Claude app, the details depend on the plan.
| Plan | Monthly price | Fable access |
|---|---|---|
| Pro | $20 ($17 a month billed yearly) | Paid with usage credits |
| Max | $100 | Included for up to half the weekly limit |
| Ultra | $200 | Included, with no five-hour limit for verified accounts |
This week only, Anthropic says, paid plans that include Fable get one to three usage resets depending on how long the account has been subscribed. Pro users and standard Team seats can also use Fable for up to half of their weekly limit during the launch week.
What Anthropic says changed
According to the company, the new versions write up to 30% faster than the 5.1 models and beat its everyday model, Claude Opus 5.5, on almost every benchmark it published, with the biggest gains in robotics simulation and remote work tasks. Anthropic also says its safety filters now block about 55% fewer legitimate requests than with Fable 5.1. These are the company's own test results and have not been checked independently.
Anthropic still recommends Opus 5.5 as the starting point for most coding and everyday tasks and Fable or Mythos only for the hardest jobs.
The release also expands an anti-copying feature to all developer accounts. It is aimed at people who try to train their own models on Claude's answers, which Anthropic says often happens through thousands of fake accounts.
