Anthropic released Claude Fable 5.2 on November 9 for demanding coding, research and long-running agent tasks. The company reports improvements in coding, reasoning and vision, with output generation up to 30% faster than Fable 5.1—not a SunMarc-tested result. Anthropic’s announcement describes the upgrade.
The practical changes
- Lower token prices: standard API rates of $8 per million input tokens and $40 per million output tokens—20% below Fable 5.1.
- Large working context: one million context tokens and up to 128,000 output tokens.
- Developer access: available as
claude-fable-5-2. Mythos 5.2 shares the underlying model but requires organizational verification; Fable is the broadly available version with additional cybersecurity and biology safeguards.
The model specifications and pricing and changes document these limits, rates and access distinctions.
Why it matters for builders
Lower token prices do not automatically mean cheaper completed tasks. Anthropic still recommends Opus 5.5 for most work, escalating to Fable when greater capability is needed. Fable can also advise smaller models or coordinate delegated work, according to Anthropic’s guidance.
For teams like SunMarc App Labs, a useful evaluation would be a bounded website migration or multi-file app fix: does it deliver a correct, reviewable result with fewer retries?
Custom agent harnesses need migration checks, too. Editing earlier conversation content can invalidate retained thinking blocks, while older computer-use integrations may require updates. The developer changes explain the affected integrations; this is not a claim that every existing workflow breaks.
For context on the suggested starting model, see our earlier coverage of Claude Opus 5.5.