Anthropic releases Opus 4.5 with new Chrome and Excel integrations

On Monday, Anthropic announced Opus 4.5, the latest mentation of its flagship model. It’s the past of Anthropic’s 4.5 bid of models to beryllium released, pursuing the motorboat of Sonnet 4.5 successful September and Haiku 4.5 successful October.

As expected, the caller mentation of Opus has state-of-the-art show connected a scope of benchmarks, including coding benchmarks (SWE-Bench and Terminal-bench), instrumentality usage (tau2-bench and MCP Atlas) and wide occupation solving (ARC-AGI 2, GPQA Diamond). 

Notably, Opus 4.5 is the archetypal exemplary to people implicit 80 percent connected SWE-Bench verified, a respected coding benchmark.

Anthropic besides emphasized the Opus’s machine usage and spreadsheet capabilities, and launched a fig of parallel products to showcase however the exemplary holds up successful those settings. Together with Opus 4.5, Anthropic volition marque its Claude for Chrome and Claude for Excel products — antecedently successful aviator — much broadly available. The Chrome hold volition beryllium disposable to each Max users, portion the Excel-focused exemplary volition beryllium disposable to Max, Team and Enterprise users.

Opus 4.5 besides comes with representation improvements for long-context operations, which required important changes successful however the exemplary manages its memory.

“There are improvements we made connected wide agelong discourse prime successful grooming with Opus 4.5, but discourse windows are not going to beryllium capable by themselves,” Dianne Na Penn, Anthropic’s caput of merchandise absorption for research, told TechCrunch. “Knowing the close details to retrieve is truly important successful complement to conscionable having a longer discourse window.”

Those changes besides enabled a long-requested “endless chat” diagnostic for paid Claude users, which volition let chats to proceed without interruption erstwhile the exemplary hits its discourse window. Instead, the exemplary volition compress its discourse representation without alerting the user.

Many of the upgrades are made with an oculus towards agentic use-cases, peculiarly scenarios successful which Opus acts arsenic a pb cause commanding a radical of Haiku-powered sub-agents. Managing those tasks requires a beardown bid of moving memory, which is wherever the representation improvements described by Penn truly amusement their worth.

“This is wherever fundamentals similar representation go truly important,” Penn says, “because Claude needs to beryllium capable to research codification bases and ample documents, and besides cognize erstwhile to backtrack and recheck something.”

Opus 4.5 volition look stiff contention from different precocious released frontier models, astir notably OpenAI’s GPT 5.1 (released connected November 12) and Google’s Gemini 3 (released November 18).