Anthropic announced the release of Opus 4.5, the latest version of its flagship large language model, on Monday. This introduction marks the final model in Anthropic's 4.5 series, following the earlier launches of Sonnet 4.5 in September and Haiku 4.5 in October. The new Opus model demonstrates state-of-the-art performance across various benchmarks and is accompanied by broader availability of its Claude for Chrome and Claude for Excel integrations.
Opus 4.5 has achieved high performance on a range of established benchmarks, including those for coding, tool use, and general problem-solving. Specifically, the model reportedly scored over 80 percent on SWE-Bench verified, a recognized coding benchmark, making it the first model to reach this threshold. Additional strong performances were noted on Terminal-bench for coding, tau2-bench and MCP Atlas for tool use, and ARC-AGI 2 and GPQA Diamond for general problem solving.
Concurrently with the Opus 4.5 release, Anthropic expanded the availability of its Claude for Chrome and Claude for Excel products, which were previously in a pilot phase. The Chrome extension is now accessible to all Max users, while the Excel-focused model has been made available to Max, Team, and Enterprise users. The company highlighted Opus's improved capabilities in computer use and spreadsheet operations, aligning with these new integrations.
The new model incorporates significant memory enhancements designed for long-context operations. These improvements have enabled a new "endless chat" feature for paid Claude users, which allows conversations to continue uninterrupted by compressing context memory without user notification. According to Dianne Na Penn, Anthropic’s head of product management for research, these memory advancements are crucial for "agentic use-cases," particularly scenarios where Opus functions as a primary agent managing a group of Haiku-powered sub-agents. Penn stated, "Knowing the right details to remember is really important in complement to just having a longer context window." She further added that "Claude needs to be able to explore code bases and large documents, and also know when to backtrack and recheck something."
The launch positions Opus 4.5 in a competitive market, facing other recently introduced frontier models. These include OpenAI’s GPT 5.1, released on November 12, and Google’s Gemini 3, which became available on November 18.
