Claude Fable 5.2: Features, Pricing and How to Use It

What is Claude Fable 5.2?

Claude Fable 5.2 is an upgraded artificial intelligence model developed by Anthropic for complex, open-ended technical work. It shares its underlying architecture with Claude Mythos 5.2 but includes built-in safety safeguards for broad public availability.

The model improves upon earlier Fable models by outputting responses up to 30% faster and offering clearer communication. Anthropic designed Claude Fable 5.2 to act as a standalone problem solver, an advisor to other models, or an orchestrator that delegates tasks to smaller AI systems.

What can Claude Fable 5.2 do?

Claude Fable 5.2 handles high-difficulty tasks across software coding, reasoning, mathematical proofs, and visual hardware evaluation. According to Anthropic, over 90% of code changes merged into their core codebase are now written by Claude models with human guidance.

During prerelease tests, Claude Fable 5.2 formalized the proof of Fermat's Last Theorem in 30 hours and designed three working electronic devices from scratch. In one experiment, an Anthropic engineer tasked the model with creating a prototype operating system, leading to a bootable OS developed by over 1,000 coordinated subagents over eight days.

A team of agents running an earlier version of Fable 5.2 also completed codebase migrations, rewriting a critical production service from Python to Rust that starts about 26 times faster. On standard evaluations, Claude Fable 5.2 scored 83.5 on BenchCAD 2.1 at Max effort and reached 49.1% on the FrontierCode 2.0 benchmark at Xhigh effort.

How much does Claude Fable 5.2 cost?

Claude Fable 5.2 costs $8 per million input tokens and $40 per million output tokens under standard operating mode. Anthropic also provides an optional fast mode that increases processing speeds at a higher fee rate.

At default Medium effort settings, Claude Fable 5.2 accomplishes challenging tasks using fewer tokens than Claude Fable 5.1. On benchmarks such as WANDR and Chartography, it matches or exceeds scores achieved by Claude Opus 5.5 for a fraction of the cost per task.

How to use Claude Fable 5.2

Claude Fable 5.2 is available to all users across Claude applications and Claude Code, running on Medium effort by default. Anthropic recommends beginning routine work in Claude Opus 5.5 and switching to Fable 5.2 depending on task scope and complexity.

Anthropic notes that Claude Fable 5.2 can operate in distinct technical roles within a workflow:

• Standalone solver for ambitious, open-ended technical challenges
• Advisor consulted by Claude Haiku, Sonnet, or Opus models to resolve difficult problems
• Orchestrator that delegates simpler tasks to smaller models

Claude Fable 5.2 vs Claude Opus 5.5 and Mythos 5.2

Claude Fable 5.2 and Claude Mythos 5.2 share the same core model, but Fable 5.2 includes public safety safeguards for general access. Claude Opus 5.5 serves as Anthropic's primary daily driver for routine technical and knowledge tasks.

While Claude Opus 5.5 remains recommended for general daily work, Claude Fable 5.2 outperforms Opus 5.5 on technical benchmarks like FrontierCode 2.0 and BenchCAD 2.1. Access to Claude Mythos 5.2 requires approval through specialized verification programs for biology or cybersecurity research.

Frequently asked questions

Sources: Anthropic