Anthropic, a startup backed by Alphabet and Amazon, has unveiled its latest AI technology, Claude Opus 4, which can autonomously write computer code for significantly longer periods than its predecessors. The new model represents a major milestone in Anthropic’s mission to develop increasingly autonomous AI systems. According to Mike Krieger, Anthropic’s Chief Product Officer, Claude Opus 4 can code autonomously for nearly seven hours, a substantial improvement over the 45 minutes achieved by its previous model, Claude 3.7 Sonnet.
This advancement is crucial for achieving the economic and productivity impact that AI is capable of, as Krieger explained in an interview with Reuters. “For AI to really have the economic and productivity impact that I think it can have, the models do need to be able to work autonomously and work coherently for that (longer) amount of time,” he said. The enhanced capability was demonstrated when customer Rakuten used Opus 4 for nearly seven hours of continuous coding, while an Anthropic researcher successfully set up the AI to play 24 hours of a Pokemon game.
In addition to Claude Opus 4, Anthropic also announced Claude Sonnet 4, a smaller and more cost-effective AI model. Both models are designed to provide flexible responses, ranging from near-instant answers to more detailed, reasoned responses over longer periods. They also feature web search capabilities, enhancing their utility across various applications.
Anthropic has been making significant strides in AI development, particularly in coding AI. The release of Claude Opus 4 and Sonnet 4 follows other major AI announcements this week, including those from Google, with which Anthropic both collaborates and competes. Furthermore, Anthropic has made its Claude Code tool generally available for software developers, following a preview in February. This tool is expected to significantly enhance developer productivity by leveraging AI for coding tasks.
