Alvin Lang
Feb 05, 2026 18:35
Claude Opus 4.6 scores 23% greater on monetary evaluation benchmarks, provides Excel and PowerPoint integrations for funding banking workflows.
Anthropic dropped Claude Opus 4.6 on February 5, positioning its newest AI mannequin as a direct play for monetary companies workflows. The headline quantity: a 23 proportion level enchancment over Claude Sonnet 4.5 on the corporate’s inner Actual-World Finance benchmark, which exams roughly 50 funding and monetary evaluation use circumstances.
The mannequin now scores 60.7% on Vals AI’s Finance Agent benchmark—a 5.47% soar from Opus 4.5—which evaluates efficiency on SEC submitting analysis. It additionally hits 76% on TaxEval, one other exterior benchmark testing tax-related reasoning.
The place Analysts Truly Work
The actual story right here is not simply benchmark scores. Anthropic is pushing Claude straight into the instruments finance professionals use each day: Excel and PowerPoint.
Claude in Excel now handles pivot tables, chart modifications, conditional formatting, and what Anthropic calls “finance-grade formatting.” The combination helps multi-file drag-and-drop and auto-compaction for lengthy conversations—addressing the copy-paste hell that plagues anybody constructing advanced monetary fashions throughout a number of tabs.
Claude in PowerPoint launches in beta for Max, Workforce, and Enterprise customers. The AI reads present layouts, fonts, and grasp slides earlier than producing new content material, theoretically letting analysts construct client-ready decks with out ranging from scratch.
The Productiveness Declare
Anthropic’s advertising supplies present side-by-side comparisons of business due diligence outputs—the form of acquisition evaluation work they are saying “would usually take a senior analyst two to a few weeks to finish.” First-pass high quality has improved noticeably, in line with companions already testing the system.
“Creating monetary PowerPoints that used to take hours now takes minutes,” stated Aabhas Sharma, CTO at Hebbia. Nico Christie, co-founder of Shortcut AI, known as it “a watershed second for spreadsheet brokers.”
Lloyd Hilton, Head of Hg Catalyst, famous the mannequin handles “unstructured knowledge and intelligently working with minimal prompting to meaningfully automate advanced evaluation.”
What’s New Below the Hood
Opus 4.6 ships with a 1-million-token context window, letting it course of large datasets in single periods. The mannequin additionally improved on BrowseComp and DeepSearchQA benchmarks, which take a look at info extraction from massive, unstructured doc units—crucial for anybody doing earnings name evaluation or regulatory submitting evaluations.
Cowork, Anthropic’s desktop app function, now lets finance groups kick off a number of analyses concurrently whereas steering Claude’s strategy on every deliverable. A company finance plugin gives pre-built workflows for journal entries, variance analyses, and reconciliation.
The Wonderful Print
Anthropic is not claiming full autonomy. “Customers ought to proceed to evaluate Claude’s outputs to make sure it meets their specs; notably for high-stakes work, human judgment stays important,” the corporate famous in its launch.
For crypto and fintech corporations evaluating AI integration, Opus 4.6 represents the clearest sign but that basis mannequin corporations are shifting past chatbot interfaces towards embedded enterprise instruments. The query now: how shortly will competing fashions from OpenAI and Google match these finance-specific capabilities?
Claude Opus 4.6 is on the market now on all paid Claude plans. The PowerPoint integration stays in analysis preview for higher-tier subscribers.
Picture supply: Shutterstock
