Anthropic launches Claude Opus 4.6 featuring new ‘agent teams’ capability
Anthropic released Claude Opus 4.6, an AI model with a 1-million-token context window and 'agent teams' for enhanced AI collaboration.Summary
Anthropic has launched Claude Opus 4.6, its latest AI model, boasting a significantly expanded 1-million-token context window and a novel “agent teams” feature. This new capability allows multiple instances of Claude to collaborate on complex tasks, with one model potentially writing code while others review, test, and document, all working in unison. The model demonstrates improved performance in coding, debugging, and software development, surpassing OpenAI’s GPT-5.2 in several benchmarks, including finance, legal, and research. Anthropic also highlights improvements in safety, with fewer refusals and reduced instances of harmful behavior. The release follows a previous rollout that impacted global tech markets and signals Anthropic’s continued progress towards achieving artificial general intelligence (AGI). A notable demonstration involved Opus 4.6, using 'agent teams', successfully building a C compiler that ran on the Linux kernel after two weeks with minimal human intervention, showcasing the potential of autonomous software development.
(Source：Firstpost)
