Anthropic reveals new Claude Fable and Mythos 5.2 models

Anthropic is launching two new artificial intelligence models named Claude Fable 5.2 and Claude Mythos 5.2.

Capabilities and Testing

The company states that both models are built to handle difficult technical tasks with better coding, reasoning, and vision than the older 5.1 versions, while also producing answers at a faster speed. Claude Fable 5.2 is built to solve tough software engineering tasks, manage agent teams, and work on long projects. Anthropic reports that the new models helped engineers finish hard software updates, design three working electronic devices from the ground up, solve two open problems involving the physics of black holes, make big progress on robotics tests, and finish the proof of Fermat's Last Theorem in 30 hours.

Availability

Claude Fable 5.2 is launching today for users. However, access to Claude Mythos 5.2 is limited only to vetted organizations at the start. These new versions come after the release of the Fable 5.1 and Mythos 5.1 models in September.

Context

Artificial intelligence companies often release updated versions of their models to offer faster speeds and better problem-solving skills for software and science work. These updates follow a steady pattern of yearly or half-yearly model releases across the tech industry. Testing these systems on hard math and physics problems helps show how well the AI can handle complex real-world logic.