Anthropic has released a preview of its new AI model, Mythos, following a major leak last week. Sensitive information about the model and its launch process was exposed without authorization due to a “misconfigured Content Management System (CMS).” Notably, the new preview will help the company see how the model works in real-world situations. A small group of partner companies will also test it to help improve cybersecurity.
Claim 30% Off TipRanks
Trade AMZN with leverageAnthropic Debuts Preview of Mythos AI
Anthropic has stated that Mythos is one of its most powerful artificial intelligence models. The preview shows that the model can scan computer systems to find and flag bugs that make software unsafe. The company claimed that Mythos has already found serious issues in different codebases, including ones that have existed for many years. This shows that the model is reliable and can detect problems that humans may miss.
The new AI model is part of an initiative called “Project Glasswin,” in which partner companies such as Microsoft (MSFT), Amazon (AMZN), and Apple (AAPL) will test it for cybersecurity work. These companies will use the model to scan their internal systems and open-source software. They will also report and share their findings across the industry to strengthen security.
Anthropic Limits Access to Keep Mythos Safe
Anthropic has stated that Mythos AI will not launch publicly yet, as only a small number of companies will be allowed to test the model to ensure its safety. This approach will reduce misuse, since the tool can be used to find and exploit system weaknesses. Limiting access will also help Anthropic monitor how the model is used.
The private company is also working with regulatory authorities to figure out how the model can be handled safely and responsibly. This is important because powerful AI tools like Mythos need to be regulated to prevent harm. This step shows that the company is highly focused on safety as it develops advanced AI tools.
What are the Best AI Stocks to Buy Right Now?
Wall Street analysts rate Nvidia (NVDA), Amazon (AMZN), and Alphabet Class A (GOOGL) as “Strong Buy,” based on TipRanks consensus data. This reflects positive sentiment toward these companies’ leadership in AI development and cloud computing. For more information on each stock’s performance, ratings, and price targets, visit TipRanks’ Stocks Comparison Center.
