Anthropic launches Cloud 4, its most powerful AI model.

Anthropic launched its new family of generative AI models, CloudOps4, on Thursday. The AI ​​startup says it is designed to demonstrate its ability to compete with leading AI assistants in a more responsible manner than its competitors.

"Cloud Ops4 is the best programming model in the world," said Anthropic CEO and co-founder Dario Amodei at the company's first developer conference.

Ops4, the startup's most powerful model, and Sonnet4 are hybrid models that can provide immediate answers or perform longer tasks in the background, such as computer programming or analyzing technical data to produce a comprehensible report.

The year 2023 marked the start of the race to launch interfaces competing with ChatGPT. Since then, AI assistants have become more capable of "thinking," taking longer to respond and displaying their thought processes, and have become increasingly autonomous.

US tech companies are currently focusing on "AI agents," tools that perform various tasks on behalf of the user.

Anthropic pioneered this field with the release of its "Computer Use" system in the fall of 2024, which allows "Cloud" to use computers like a human: browsing the internet, selecting buttons on a website, entering text, and using various programs.

"Agents can transform human imagination into reality on an unprecedented scale, and this is especially important for developers like you," said Mike Krieger, Anthropic's chief product officer and Instagram co-founder.

He recalled the early days of Instagram before its acquisition by Facebook, when its small team had to choose between developing a new feature or improving the mobile app, for example.

He added Krieger: “Thanks to AI agents, startups can now run parallel experiments, improve with user feedback, and build products faster than ever before.”