Anthropic releases Claude Fable 5.2 with lower prices and a new fast mode

Anthropic has released Claude Fable 5.2, the newest version of its top-tier model family. It launched on November 9, 2026, alongside a restricted sibling called Claude Mythos 5.2.

The headline numbers are a performance gain of up to 30% over Fable 5.1 and a lower price.

What changed in Fable 5.2

Pricing is the clearest change. Fable 5.2 costs $8 per million input tokens and $40 per million output tokens.

The previous rates were $10 for input and $50 for output. That works out to a 20% cut on both sides of the meter.

On performance, the research describes an improvement of up to 30% over Fable 5.1. The key word is “up to,” which in product launches tends to describe the best case rather than the typical Tuesday.

Fable 5.1 itself only arrived two months earlier, in September 2026.

The update also adds a fast mode. It delivers output 2.5 times faster, but it costs double the standard price.

By default, Fable 5.2 runs in a high-effort mode built for complex tasks. Anthropic also says the safety features apply across the board rather than being bolted on to certain modes.

The model targets technical and knowledge-heavy work. Coding and enterprise documentation are among the use cases highlighted for the release.

Fable 5.2 is available on the Claude Platform and related applications. Before launch, third-party institutions including the AI Security Institute evaluated the model and confirmed its advancements.

A bumpy road to version 5.2

The Fable line sits at the top of Anthropic’s model lineup. The company designates the family as “Mythos-class” and pitches it for long-running, intricate projects.

Its debut was not exactly smooth. Fable 5 launched in June 2026, then spent approximately 18 days suspended because of US export restrictions.

The model was reinstated on July 1.

Fable 5 was meant to bring advanced capabilities to a wider audience. Before that, those capabilities had been limited to select partners through an initiative called Project Glasswing.

Fable 5.1 followed in September with improvements geared toward research applications. Version 5.2 now shifts the emphasis toward efficiency and cost.

The broader fifth generation of Claude models had already posted strong benchmark results. Prior metrics included a score of 80.3% on SWE-bench Pro, a test of how well models handle real software engineering tasks.

Mythos 5.2 remains a restricted variant. Anthropic is releasing it alongside Fable 5.2, but it is not getting the same broad availability.