← Back
OpenAI
OpenAI launches ChatGPT for Excel and GPT-5.4; achieves 87.3% on financial modeling benchmarks
OpenAI APIChatGPTOpenAI · releasefeaturemodelintegrationapi · openai.com ↗

ChatGPT for Excel: Direct Spreadsheet Integration

OpenAI introduced ChatGPT for Excel in beta, a native Excel add-in that brings AI-powered spreadsheet modeling directly into workbooks. Users can describe tasks in plain language, and ChatGPT will create or update live models, run scenario analysis, and generate formatted outputs—all while preserving Excel-native structures, formulas, and assumptions.

Key capabilities include:

  • Model building and updates: Teams can request ChatGPT to build financial models, run data analysis, manage reporting, and conduct budgeting using natural language
  • Spreadsheet reasoning: ChatGPT understands connections across sheets, explains output changes, traces errors, and shows how assumptions flow through models
  • Audit and control: All calculations run directly in Excel with cell-by-cell references and explanations; ChatGPT requests permission before making changes

ChatGPT for Excel is available today for Business, Enterprise, Edu, Teachers, Pro, and Plus users in the U.S., Canada, and Australia. Google Sheets support is coming soon.

GPT-5.4: Finance-Optimized Model with Significant Performance Gains

OpenAI released GPT-5.4 (Thinking) across ChatGPT, Codex, and the API, with performance specifically tuned for financial workflows. On OpenAI's internal investment banking benchmark, GPT-5.4 Thinking achieved 87.3% compared to 43.7% for GPT-5—a 2x improvement on real-world tasks including financial modeling, scenario analysis, and long-form research.

Financial Data Integrations and Research Improvements

ChatGPT now includes integrations with major financial data providers:

  • Available today: Moody's, Dow Jones Factiva, MSCI, Third Bridge, MT Newswire
  • Coming soon: FactSet, and others

Users can also build custom integrations using OpenAI's Model Context Protocol (MCP) to connect proprietary internal data. The research feature has been enhanced with:

  • Source-specific search options to focus on particular websites or databases
  • Greater control over research planning and iteration
  • Redesigned workspace for reviewing sources and citations
  • PDF and Microsoft Word export for structured, cited outputs

These features address key pain points in financial analysis: reducing manual data gathering, enabling faster due diligence, and producing audit-ready, cited outputs for valuation, underwriting, and research work.