The technology company Anthropic is bringing new tools to the public with its latest release. In a major update for developers and researchers, Anthropic launches Claude Fable 5.2 and Mythos 5.2 to help solve complex problems faster than before. These new tools bring huge improvements to coding, math, and science tasks. People can expect quicker response times and lower costs, making it easier to build new software and devices from scratch without spending too much money or waiting for long processing times.
New models bring speed and safety to coding projects
The company built these new systems to handle big technical jobs on their own. Now that Anthropic launches Claude Fable 5.2 and Mythos 5.2, users get systems that are thirty percent faster and twenty percent cheaper than older versions. They work well for creating software or testing designs for physical products.
For example, engineers used a large group of these digital helpers to create a working computer operating system in just eight days. The system even made a login screen, storage, and an installer from scratch. They also helped rewrite old code into a new programming language, making a service run twenty-six times faster.
Don’t miss the best of The Mac Observer
Set us as a preferred source and our Apple reporting ranks higher in your Google Search results and Discover feed — one tap, no account changes.
While both versions share the same core design, they serve different groups. Fable is available to everyone and comes with strong safety blocks to stop people from misusing it for computer attacks or biology risks. It blocks many risky requests to keep the general public safe while they build everyday AI applications.
Special versions help approved groups find software weaknesses
Mythos works without those tight limits, but it is only for trusted groups working in security or life sciences. Organizations must pass a verification program to get access to it. This keeps powerful artificial intelligence out of the hands of people who might use it poorly.
In testing, this special version found more software weaknesses than previous models. It helps researchers spot security gaps and even predicts the shapes of virus proteins. This makes it a great tool for people trying to defend computer networks or study complex biology.
Users will pay eight dollars for every million input words and forty dollars for every million output words, which is a good deal for the performance. The company also offers a faster mode for those willing to pay higher prices for quick results.
These tools represent a big step forward in what machines can do without constant human help. The company continues to push limits while trying to keep things secure. This launch gives builders better options to create fast, smart programs that solve real problems on a daily basis.
Discussion