2 min read
On Monday, November 24, Anthropic officially announced the release of Opus 4.5. This is the latest version of the company’s flagship AI model. With the new release, it marks the completion of the 4.5 series, following the launch of Sonnet 4.5 in September and Haiku 4.5 in October.
Opus 4.5 arrives during a busy month for AI advancements. It faces stiff competition from OpenAI’s GPT 5.1, released on November 12, and Google’s Gemini 3, which launched on November 18.
State-of-the-Art Performance
The new Opus model delivers state-of-the-art performance across a range of benchmarks. Notably, Opus 4.5 is the first model to score over 80% on the SWE-Bench verified. This is a highly respected coding benchmark.
In addition to coding success on SWE-Bench and Terminal-bench, the model excels in other areas. It shows top-tier results in tool use benchmarks like tau2-bench and MCP Atlas. It also demonstrates strong general problem-solving capabilities on ARC-AGI 2 and GPQA Diamond.
New Chrome & Excel Integrations
Anthropic is also expanding how users interact with the model. Alongside the Opus 4.5 launch, the company is making its “Claude for Chrome” and “Claude for Excel” products more broadly available. Previously, these tools were only in pilot programs.
The Chrome extension is now available to all Max users. Meanwhile, the Excel-focused model is available to Max, Team, and Enterprise users. These products showcase the model’s improved computer use and spreadsheet capabilities.
Anthropic Brings Memory Improvements & Endless Chat
Opus 4.5 introduces significant changes to how the model manages memory. These upgrades enable a long-requested “endless chat” feature for paid users. Consequently, chats can now proceed without interruption. When the model hits its context limit, it compresses its memory without alerting the user.
Dianne Na Penn, Anthropic’s head of product management for research, emphasised the importance of this update. Penn stated:
Knowing the right details to remember is really important in complement to just having a longer context window.
These memory improvements are vital for agentic use cases. For example, Opus can act as a lead agent commanding a group of Haiku-powered sub-agents. According to Penn, these fundamentals allow Claude to explore code bases and large documents while knowing when to backtrack and recheck information.
