Anthropic releases Claude Fable 5.2 and Claude Mythos 5.2

What shipped on November 9

Fable 5.2 and Mythos 5.2 are one model with two sets of safeguards, the same arrangement Anthropic used for the 5.1 release. Fable 5.2 went out on Anthropic's API and through Amazon Web Services, Google Cloud and Microsoft Azure. Mythos 5.2 is limited to organizations admitted through Anthropic's Cyber and Life Sciences Verification Programs. [1][2]

Both cost $8 per million input tokens and $40 per million output tokens, with a fast mode at $16 and $80. The 5.1 models were priced at $10 and $50, so the per-token price fell 20 percent. Anthropic says outputs arrive up to 30 percent faster than with 5.1. [1][3]

Anthropic's capability claims

On the Remote Labor Index, which scores models on paid freelance projects, Anthropic reports 41.3 percent for Fable 5.2 at high effort, against 17.9 percent for Fable 5.1 and 21.3 percent for Opus 5.5. On its CVE Rediscovery evaluation, Anthropic says Mythos 5.2 found 10.2 percent more vulnerabilities than Opus 5.5 and 14.4 percent more than Mythos 5.1. [1]

The announcement describes two long-running demonstrations. In one, an Anthropic engineer had Fable 5.2 coordinate more than 1,000 subagents for eight days to build a prototype operating system that boots on a laptop. In the other, 1,000 Fable 5.2 agents formalized Fermat's Last Theorem in 30 hours; Anthropic says an earlier unreleased research model took 11 days, and Opus 5.5 completed 26 percent in 18 hours. [1]

What the evidence does not establish

Every figure above comes from Anthropic, measured on Anthropic's own harnesses. The operating system and the formalization are single runs described by the company, without an outside account of how much human steering they needed or how complete the results are. The same-day coverage repeats the figures rather than testing them. [1][2]