Get ready to be amazed—Anthropic just dropped Opus 4.5, and it’s a game-changer in the world of AI models. But here’s where it gets controversial: while some hail it as a breakthrough, others are already questioning whether it can truly outshine competitors like OpenAI’s GPT 5.1 and Google’s Gemini 3. Let’s dive in.
On Monday, Anthropic unveiled Opus 4.5, the latest and final addition to its 4.5 series, following the releases of Sonnet 4.5 in September and Haiku 4.5 in October. As expected, Opus 4.5 delivers cutting-edge performance across a range of benchmarks, from coding (SWE-Bench and Terminal-bench) to tool use (tau2-bench and MCP Atlas) and general problem-solving (ARC-AGI 2, GPQA Diamond). And this is the part most people miss: it’s the first model to surpass 80% on SWE-Bench verified, a benchmark that’s notoriously tough to crack.
What sets Opus 4.5 apart isn’t just its scores—it’s also its enhanced computer and spreadsheet capabilities. Anthropic is doubling down on this by expanding access to Claude for Chrome and Claude for Excel, tools previously in pilot. The Chrome extension is now available to all Max users, while the Excel-focused model is open to Max, Team, and Enterprise users. This move signals Anthropic’s commitment to making AI more practical for everyday tasks.
Here’s the kicker: Opus 4.5 introduces significant memory improvements for long-context operations, a feature that’s been a long-standing challenge in AI. Dianne Na Penn, Anthropic’s head of product management for research, told TechCrunch, “Knowing the right details to remember is just as crucial as having a longer context window.” These upgrades also enable a highly anticipated ‘endless chat’ feature for paid Claude users, allowing conversations to flow uninterrupted even when the model hits its context limit. Instead of stopping, the model compresses its memory seamlessly—no user intervention required.
Many of these enhancements are tailored for agentic use cases, particularly scenarios where Opus acts as a lead agent managing Haiku-powered sub-agents. This requires a robust working memory, and that’s where Opus 4.5’s improvements truly shine. As Penn explains, “Claude needs to navigate complex codebases and documents while knowing when to backtrack—memory is key.”
But Opus 4.5 isn’t entering the arena unopposed. It faces stiff competition from OpenAI’s GPT 5.1 (released November 12) and Google’s Gemini 3 (released November 18). Here’s the question we’re all asking: Can Opus 4.5 hold its ground against these heavyweights? Only time will tell.
For those eager to explore, Claude for Chrome is available at https://www.claude.com/blog/claude-for-chrome, and Claude for Excel at https://www.claude.com/claude-for-excel. And if you’re curious about the author behind this story, Russell Brandom has been covering the tech industry since 2012, with a focus on platform policy and emerging technologies. Reach out to him at russell.brandom@techcrunch.com or on Signal at 412-401-5489. View his full bio here.
Now, we want to hear from you: Do you think Opus 4.5 is a game-changer, or is it just another player in an already crowded field? Let us know in the comments!
