Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Anthropic’s “smarter model” is getting a big boost, the company said in a blog post announcing Claude Opus 4.6.
He described the new model as a “direct upgrade” from its predecessor in a release, noting that it can better handle complex, multi-step tasks and gets much closer to production-ready quality on the first try than we’ve seen with any model — documents, spreadsheets, and presentations will need less iteration. It is available starting today at the same price as the previous version, and according to the company, its particular strengths lie in proxy coding, use of tools, and financial research and analysis.
But most of all, it seems that with this release, Anthropic wants to expand the current hype for Claude beyond just programming and corner the market on other types of knowledge work. With Opus 4.6, you invested in improving the model in creating presentations in PowerPoint and documents in Excel. The blog post included a plugin for Cowork, Anthropic’s latest release that’s a tech-friendly version of Claude Code, with the hope that users in non-tech industries will explore use cases for research, marketing, and more.
On the programming front, Anthropic said in a statement that Opus 4.6 was designed to further improve developers’ experience with Claude Code, since it specializes in long-running tasks and can “take a development project that would normally take days and finish it in hours, handling everything from architecture to deployment.”
The company also announced a feature currently in research preview called “Agent Teams,” allowing the new model to work within Claude Code “the way a real engineering team would,” meaning it’s possible to split the work of a single project across agents who each own a part of the project and coordinate with each other.
Diane Na Bin, Head of Research Product Management at Anthropic, said: Edge The company is focused on improving the “multi-agent” experience for developers with this launch, investing in the quality and speed of output, as well as making the model better at other types of knowledge work besides just programming – such as Excel, PowerPoint, and search functions.
“This is the first version of the Opus model where we have a contextual window with 1 million views in the beta,” Ben said. “We’ve just received positive feedback about Opus 4.5 where one of the main features people wanted was a longer context window so they can work with Claude across more documents.”
Anthropic said in the blog post that it conducted the “most comprehensive” set of safety tests for the Opus 4.6 of any of its models to date. The new evaluations included assessments of user well-being, more sophisticated tests on whether Opus 4.6 could reject “potentially dangerous requests,” and updated tests of the extent to which the model could covertly perform malicious actions. The model also showcases heightened cybersecurity capabilities, according to the company, so it includes six new cybersecurity probes to track potential misuse.