Kamal Ghorui
Anthropic, the AI startup whose workplace tool triggered a massive sell-off in software stocks earlier this week, has launched Claude Opus 4.6—its most powerful model yet. The timing couldn’t be more loaded.The new model, announced Thursday, brings stronger coding abilities, a 1 million token context window (up from 200,000), and a feature called “agent teams” that lets multiple AI agents split tasks and work in parallel. Think of it as a virtual engineering squad that coordinates on its own. Opus 4.6 also supports outputs of up to 128,000 tokens, meaning it can handle much larger tasks without needing to break them into multiple requests.On benchmarks, the model outperforms OpenAI’s GPT-5.2 on GDPval-AA, an evaluation of real-world knowledge work in finance, legal and other professional domains—precisely the sectors that just saw their stocks crater because of Anthropic’s earlier product update. It also tops every other frontier model on BrowseComp, which measures a model’s ability to dig up hard-to-find information online.
A legal plugin that spooked Wall Street
Here’s the backstory. Last Friday, Anthropic released a set of open-source plugins for Claude Cowork, its AI assistant built for non-coders. Most covered standard business functions like sales and marketing. But one plugin, designed to automate contract reviews, NDA triage and legal briefings, sent shockwaves through the market.Thomson Reuters fell nearly 16% on Tuesday. LegalZoom crashed close to 20%. A Goldman Sachs basket of US software stocks dropped 6% in a single session—the worst since April. By the time the dust began to settle on Wednesday, roughly $830 billion had been wiped from the S&P 500 software and services index since January 28. Traders started calling it the “SaaSpocalypse.”The fear wasn’t just about legal tech. Investors saw Anthropic muscling into the enterprise application layer, threatening the subscription-based business models that power companies like Salesforce, Intuit and Adobe. Private equity firms with heavy software portfolios also took hits—Ares Management, KKR and Blue Owl Capital all dropped more than 9% in a single day.
Opus 4.6 doubles down on exactly what spooked investors
Now, with Opus 4.6, Anthropic is making its AI even better at the kind of work that triggered the panic. The model produces documents, spreadsheets and presentations that are closer to “production-ready” on the first try, the company says. It handles financial modelling, regulatory filings and long-form research with less hand-holding.Harvey, the legal AI startup, reported that Opus 4.6 scored 90.2% on its BigLaw Bench evaluation—the highest of any Claude model. Box said its testing showed a 10% lift in performance on multi-source analysis across legal, financial and technical content.Anthropic also introduced Claude in PowerPoint as a research preview, letting users build and edit slide decks directly within the app. The AI reads layouts, fonts and slide masters to stay on brand—a feature available on Max, Team and Enterprise plans.Scott White, Anthropic’s head of product, told CNBC that the company is entering a “vibe working” era, where non-developers can hand real, significant work to AI. Enterprise customers already make up roughly 80% of Anthropic’s business.
Not everyone’s buying the doom narrative
Not everyone thinks the sky is falling. Nvidia CEO Jensen Huang called the software sell-off “the most illogical thing in the world,” arguing AI will use existing tools rather than replace them. Google’s Sundar Pichai struck a similar tone during his company’s earnings call, saying companies that “seize the moment” will find the same opportunities ahead.But the market isn’t fully convinced. Claude Opus 4.6 is available today on claude.ai, Anthropic’s API and all major cloud platforms. Pricing stays at $5/$25 per million tokens. Whether it calms investors or spooks them further is the real test.
