Anthropic announced that it is working on a new model family with the Claude Opus 4.8 model. The company stated that Mythos-level models, which will exceed the performance limits of Opus, will soon be available to users.

Currently, a limited number of organizations have access to the Claude Mythos Preview model as part of the Glasswing Project. This particular model is especially optimized for work in the field of cyber security.

New model family is on the way

Anthropic officials emphasize that the Mythos model must pass strict security tests before being made available for general use. The company is working hard to develop these defense systems.

The phrase ClaudeMythos-5 detected in the developer mode shows that the new model will be positioned as a different class alongside the Haiku, Sonnet and Opus series. This naming indicates that more than one version may be released under the umbrella of Mythos.

This powerful model, which is planned to be made publicly available in the coming weeks, aims to reinforce Anthropic’s thesis in the technology world. This breakthrough is an important technological show of strength for the company, which is preparing for the public offering process.

Cybersecurity focused performance

Claude Mythos is designed to perform more complex tasks, exceeding the capabilities offered by existing models. This new model is expected to set standards, especially in areas that require high sensitivity, such as cyber security.

The company states that they are making rapid progress in security measures and that if everything goes as planned, users will reach this performance in a short time. We will see together how this new model, which the technology world is eagerly awaiting, will change the competitive stability in the market.

What difference do you think Claude Mythos will make among existing artificial intelligence models? Could you share your intentions with us?