Anthropic Claude Fable 5.2

Anthropic released Claude Fable 5.2 and Claude Mythos 5.2 on November 9, 2026 as updated models for complex programming, scientific research, and autonomous agent operation, claiming up to a 30% generation speedup and a 20% API cost reduction, with an optional Fast mode running up to 2.5x faster. The model can also coordinate hundreds to thousands of auxiliary agents, checking results and combining them into a finished solution. According to Anthropic's published testing, Fable 5.2 improved RoboDojo Sim from 27.6% to 64.9%, the Remote Labor Index from 17.9% to 41.3%, BenchCAD 2.1 from 44.4 to 83.5 points, and FrontierCode 2.0 from 32.4% to 49.1% versus Fable 5.1. The report also says Fable 5.2 coordinated over 1,000 agents to build an operating-system prototype with kernel, drivers, file system, and GUI in eight days, and produced two new results on black hole physics and quantum gravity published as preprints.

DataCamp compares Anthropic's Claude Fable 5.2, released to all users in November 2026, against OpenAI's GPT-6 Astra across long-horizon knowledge work, coding, computer use, safeguards, and pricing, and runs one hands-on build task on both models. The author notes Fable 5.2 ships with a price cut relative to Fable 5.1 and that Anthropic's launch table places it ahead of Astra on every shared row. The article reports Fable 5.2 is cheaper per token and much cheaper for cached prompts and for requests above Astra's long-context threshold, though in the hands-on shader test it produced the better scene at higher run cost because it consumed far more tokens. The author cautions that every head-to-head score comes from Anthropic, since OpenAI published no comparison against Fable 5.2, so the benchmark gap remains Anthropic's case until independent runs appear.

Anthropic published the system card for Claude Fable 5.2 and Claude Mythos 5.2 on November 9, 2026, describing them as two configurations of the same new large language model. Fable 5.2 is released for general use with added safeguards blocking certain tasks in high-risk dual-use areas such as biology and cybersecurity, while Mythos 5.2 carries more permissive safeguards in those domains. Mythos 5.2 access is restricted to vetted individuals and organizations through trusted access programs, and it also powers Claude Security for Claude Enterprise customers. Under Anthropic's Responsible Scaling Policy, autonomy threat model 1 applies to Mythos 5.2 but remains at the low risk level assessed in the August 2026 Risk Report, and chemical/biological evaluations did not cross the CB-2 threshold, so CB-1 protections apply. On cybersecurity, Mythos 5.2 met or exceeded Mythos 5.1 and Opus 5.5.