No, Anthropic’s new Claude Opus 4.7 model is not a Mythos preview

Anthropic on Thursday released a new AI model, and no, it’s not the Claude Mythos Preview. Claude Opus 4.7 Now generally available, it aims to help developers and programmers with the most difficult programming tasks.

Opus 4.7, like a well-trained dog, is supposedly better at following instructions. Anthropic wrote in its blog post that Opus 4.7 takes the instructions “literally,” where previous models skipped the instructions or interpreted them loosely. It has improvements over the file-based memory system, so it should be able to recall information from previous sessions and documents. It can handle larger image files and analyze data from charts more easily.

Anthropic also said that the model is more “tasteful and creative” when creating interfaces, documents, and slide decks. There are no details about what anthropologists consider bad taste versus good taste.

Anthropic made waves earlier this month when it revealed that it had created a Claude Mythos Preview model, a next-generation model, but the model was… Very good at finding vulnerabilities That the company will share it with technology and Internet infrastructure companies — such as Cisco, CrowdStrike, and Amazon Web Services — so they can address the issues found by Mythos.

The idea is that if technology companies can improve their systems with the help of AI, they will be more resilient to cyberattacks by bad actors who can use publicly available AI models like everyone else.

Although Opus 4.7 is not the same as Mythos, Anthropic is testing some new cybersecurity protections in Opus 4.7. These safeguards, which “automatically detect and block requests that indicate prohibited or high-risk cybersecurity uses,” are the watered-down version of what will be in the “Mythos Class” models, the company’s blog post said. But they remain important as cybersecurity becomes increasingly infused with artificial intelligence, both for defense and for attack.