Anthropic Unveils Opus 4.5: Enhanced Chrome and Excel Integrations
On Monday, Anthropic introduced Opus 4.5, the latest addition to its model lineup, marking the conclusion of its 4.5 series. This follows the previous releases of Sonnet 4.5 in September and Haiku 4.5 in October.
Opus 4.5 showcases top-tier performance across various benchmarks, excelling in coding assessments such as SWE-Bench and Terminal-bench, as well as tool usage tests like tau2-bench and MCP Atlas. Additionally, it demonstrates prowess in general problem-solving challenges, evidenced by its scores on ARC-AGI 2 and GPQA Diamond.
A standout feature of Opus 4.5 is its achievement of over 80 percent on SWE-Bench verified, a benchmark recognized in the coding community. Anthropic has also highlighted improvements in the model’s computer usage and spreadsheet capabilities. Alongside Opus 4.5, the company is expanding access to Claude for Chrome and Claude for Excel, previously available only in pilot. The Chrome extension will now be accessible to all Max users, while the Excel integration will serve Max, Team, and Enterprise users.
Significant memory enhancements in Opus 4.5 support long-context operations, necessitating substantial adjustments in memory management. One of the key upgrades is the highly anticipated “endless chat” feature for paid Claude users, enabling uninterrupted conversations even as the model approaches its context window. Instead of notifying users when it reaches this limit, Opus 4.5 will intelligently compress its context memory.
Many of the updates are designed with agentic use-cases in mind, particularly scenarios where Opus functions as a lead agent directing a team of Haiku-powered sub-agents. Managing these tasks requires robust working memory capabilities, where the recently improved memory management shines.
Penn highlights this importance: “This is where fundamentals like memory become really important, because Claude needs to explore code bases and large documents, while also knowing when to backtrack and recheck something.”
As Opus 4.5 enters the market, it will contend with other leading models, such as OpenAI’s GPT 5.1, released on November 12, and Google’s Gemini 3, which debuted shortly after on November 18.
