OpenAI launches GPT-5.4 with native computer use mode, financial plugins for Microsoft Excel, Google Sheets

· Source: VentureBeat · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Data Science & Analytics · Depth: Advanced, medium

Summary

OpenAI has launched GPT-5.4, an advanced AI model available in two versions: GPT-5.4 Thinking and GPT-5.4 Pro. This release, following GPT-5.3 Instant, introduces a "native" Computer Use mode via API and Codex, allowing the model to navigate and operate across computer applications. Key features include new integrations for Microsoft Excel and Google Sheets, enabling granular financial analysis and automated tasks. OpenAI reports significant efficiency gains, with GPT-5.4 using 47% fewer tokens on some tasks, and improved factual accuracy, reducing false claims by 33%. The model supports up to 1 million tokens of context, though costs double for inputs exceeding 272,000 tokens. Benchmarks like OSWorld-Verified show GPT-5.4 Pro achieving 89.3% success in web browsing and 75.0% in desktop navigation, surpassing human performance in some areas.

Key takeaway

For CTOs and VPs of Engineering evaluating AI for enterprise automation, GPT-5.4's native computer use and deep financial integrations present a compelling case for deploying agentic systems. Your teams can now build AI solutions that operate across applications and spreadsheets, potentially streamlining complex workflows and reducing manual effort, but be mindful of the higher token costs for very long contexts.

Key insights

GPT-5.4 introduces native computer control and deep spreadsheet integration, advancing AI towards autonomous, multi-step professional workflows.

Principles

Method

GPT-5.4 employs tool search to dynamically retrieve tool definitions, reducing context pollution and token usage. It also uses code generation (Playwright) and direct UI interaction (mouse/keyboard via screenshots) for computer control.

In practice

Topics

Best for: CTO, VP of Engineering/Data, Director of AI/ML, AI Engineer, MLOps Engineer, AI Product Manager

Related on AIssential

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by VentureBeat.