Anthropic introduced Claude Fable 5.2 and Mythos 5.2: what the new AI models can do

Anthropic announced two new models of artificial intelligence: Claude Fable 5.2 and Claude Mythos 5.2. The company stated that the tools were developed to solve complex technical challenges and improve performance in programming, reasoning, and visual processing.

The models can also generate responses faster than their 5.1 versions. While Fable 5.2 is already available, initial access to Mythos 5.2 will be limited to organizations previously evaluated by Anthropic.

What Claude Fable 5.2 and Claude Mythos 5.2 Can Do

Anthropic highlighted the capabilities of the new models to tackle software engineering problems, coordinate teams of artificial intelligence agents, and develop projects that require work over extended periods.

In particular, Claude Fable 5.2 is geared towards complex programming challenges and has also shown results in scientific research. The company claimed that the model managed to solve two theoretical physics problems that remained open.

During pre-launch testing, Anthropic also used both models for various technical tasks. Among the results communicated by the company are:

• Completing software migrations that their engineers considered unfeasible.

• Designing three functional electronic devices from scratch.

These results were reported by Anthropic as part of the tests conducted before the launch of the models.

When Claude Fable 5.2 and Mythos 5.2 Can Be Used

The availability of the two models will be different. Claude Fable 5.2 has been launched and is already available, while Claude Mythos 5.2 will initially have restricted access.

In the case of Mythos, Anthropic will enable its use for previously evaluated organizations. The company did not detail in the announcement when this availability will be expanded.

What Changes from Claude Fable 5.1 and Mythos 5.1

The new models arrive after the launch of Claude Fable 5.1 and Claude Mythos 5.1, presented by Anthropic in September.

The update to version 5.2 incorporates improvements in programming, reasoning, and vision, as well as increased speed in generating results. With these changes, the company aims to expand the capabilities of its models to solve more complex technical and scientific problems.