Anthropic Launches Claude Fable 5.2 With Coding Gains

Anthropic has launched Claude Fable 5.2, a new broadly available AI model that the company says delivers major gains in coding, reasoning and vision work. Announced on 9 November, Fable 5.2 replaces Fable 5.1 in Anthropic’s current model tier and is available immediately through Claude’s paid plans, its API and several cloud platforms.

The company also announced Claude Mythos 5.2. It uses the same underlying model as Fable 5.2, but has fewer restrictions and is only available to organisations approved through Anthropic’s verification programmes, including its cybersecurity and life-sciences schemes.

Faster coding claims and lower API prices

Anthropic says Fable 5.2 can generate output up to 30% faster than Fable 5.1. On its FrontierCode 2.0 benchmark, which measures whether proposed code changes would be accepted by maintainers of open-source repositories, Fable 5.2 scored 49.1% at its Xhigh effort setting, up from 32.4% for Fable 5.1.

The company also reported a 69.9% score on FrontierSWE v2, versus 56.3% for its predecessor. Anthropic says more than 90% of changes merged into its own core codebase are now written by Claude with human guidance. It also says an earlier Fable 5.2 version helped migrate a production service from Python to Rust, making it approximately 26 times faster.

Those figures are Anthropic’s own results. Independent confirmation of its reported scientific and engineering achievements was not confirmed in the sources checked.

“Claude Fable 5.2 and Claude Mythos 5.2 show improvements in areas like coding, reasoning, and vision.”

A million-token context window

Both new models have a one-million-token context window and can produce outputs of up to 128,000 tokens. Adaptive thinking is always enabled, with five effort levels: low, medium, high, xhigh and max.

Standard API pricing is $8 per million input tokens and $40 per million output tokens, down from $10 and $50 respectively for Fable 5.1. Anthropic also offers an optional Fast mode at $16 per million input tokens and $80 per million output tokens, which it says can increase generation speed by up to 2.5 times.

A separate UK price has not been confirmed. Anthropic’s API documentation says billing is in US dollars rather than pounds.

Who can use it?

Fable 5.2 is available through Claude applications, the Claude Platform, Amazon Bedrock, Google Cloud, Microsoft Foundry and Claude Code. For consumer users, Anthropic lists it on paid Pro, Max, Ultra, Team and Enterprise plans; it is not available on the Free plan. Some plans include usage within their limits, while Pro and certain Team or Enterprise arrangements may require additional usage credits.

Mythos 5.2 is not a consumer product. Its availability is limited to approved partners and organisations that pass Anthropic’s verification requirements.

Anthropic says the covered models have a 30-day minimum data-retention period by default. Eligible organisations may obtain zero-data-retention arrangements through Enterprise Frontier Safeguards or other authorised arrangements, but that is not the standard setup for ordinary users.

Why it matters

Fable 5.2’s combination of lower standard API costs, a much larger context window and claimed coding improvements could make it more attractive for UK developers using Claude on complex software projects. But the most striking performance claims are still vendor-reported, and access is split sharply: Fable is aimed at paid customers, while the less restricted Mythos version remains for vetted organisations.

Anthropic acknowledges the limits of its safety testing, writing: “The problem of building evaluations that reliably catch every failure before a model is deployed remains unsolved.”