Claude Opus 4.6 arrives, Anthropic's artificial intelligence targets business
Extended context, collaborative agents and integration into business flows: the challenge to OpenAI and Google is not on show, but on real business productivity.
After the commercials to mock rival OpenAI Anthropic, Anthropic is making a comeback. And it does so with a clear message: artificial intelligence should no longer just talk well, but do useful things all the way. Claude Opus 4.6 announced yesterday is an update of its smartest model: it presents itself not as a smarter chatbot, but as a decisive step towards AI as a digital worker. The novelty? The new artificial intelligence can work on tasks longer and more reliably, while showing improvements in programming and finance, Anthropic said.
The move comes at a delicate moment. The market for language models is crowded, noisy, hyper-competitive. On the one hand, OpenAI is pushing for mass adoption. On the other, Google is trying to bring Gemini everywhere, from smartphones to the cloud, integrated into its applications. Anthropic chooses a third way: after becoming the most popular LLM among programmers and computer scientists, it wants to attack the market for intelligent agents.
The core of the model is the ability to handle huge contexts, up to a million tokens in beta (like its rival Gemini). Translated: it can read entire archives, kilometre-long reports, code repositories, and keep them all in its head while it works. It's like going from a desk to an open space.
But the real discontinuity is another. Claude no longer thinks alone. He works in a team. This AI, in the computer programming tool Claude Code, promises to divide tasks between several autonomous agents and get the job done faster. Anthropic in this way introduces so-called collaborative agents: multiple instances of the model that divide up tasks, monitor each other, and advance by objectives. It is the idea of AI as a project team, not as a single virtual consultant. One model plans, another executes, a third verifies. Less prompting, more results.
This makes Claude Opus 4.6 particularly suitable for corporate work. It doesn't just suggest a presentation: it builds it. It does not propose a strategy: it develops it, documents it, refines it. Anthropic speaks openly about end-to-end tasks, from raw input to board-ready output. It is a paradigm shift: from assistance to execution.
