Anthropic released Claude Opus 4.6 on Thursday, positioning its most advanced artificial intelligence model as a tool for knowledge workers beyond software engineering whilst markets continue to punish traditional software stocks.
The San Francisco-based company, backed by Amazon and Alphabet’s Google, said Opus 4.6 excels at sustained complex tasks, coding, financial analysis and document creation.
“We are the leader in the enterprise market and we want to be the ones building real-life agents for the enterprise,” Guillaume Princen, Anthropic’s head of digital native businesses, told the Financial Times.
The launch arrives as Anthropic’s tools have rattled technology stocks. The WisdomTree Cloud Computing Fund has fallen more than 20 per cent year to date, whilst Salesforce, Workday and Thomson Reuters shares each traded around 3 per cent lower on Thursday, Reuters reported.
Opus 4.6 outperformed OpenAI’s GPT-5.2 on an independent benchmark measuring performance on knowledge-work tasks across finance and legal sectors, the company said. The model can now process 1 million tokens in a single prompt.
Anthropic introduced “agent teams” in Claude Code, allowing multiple agents to work in parallel on complex projects. The company also expanded capabilities in Microsoft Office applications, with improved Excel functionality and a new PowerPoint integration.
“If I think about the last year, Claude went from a model that you can sort of talk to to accomplish a very small task or get an answer, to something that you can actually hand real significant work to,” Scott White, Anthropic’s head of product for enterprise, told CNBC.
Before launch, Anthropic’s frontier red team tested Opus 4.6 in a sandboxed environment, where the model identified more than 500 previously unknown high-severity security flaws in open-source libraries, Axios reported. Each vulnerability was validated by either an Anthropic team member or an outside security researcher.
“It’s a race between defenders and attackers, and we want to put the tools in the hands of defenders as fast as possible,” Logan Graham, head of Anthropic’s frontier red team, told Axios.
The company has implemented new security controls to identify and respond to potential misuse of the model’s enhanced cyber capabilities, including possible real-time detection tools.
Anthropic is in talks to raise about $20 billion from venture capitalists and other investors, double its initial target, the Financial Times reported. The funding would value the company at $350 billion ahead of its highly anticipated initial public offering.
