OpenAI introduces GPT-5.4 with more knowledge-work capability

2026-03-05 · Source: AI - Ars Technica · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Intermediate, short

Summary

OpenAI has released GPT-5.4, including GPT-5.4 Thinking and GPT-5.4 Pro, amidst user attrition to competitors like Anthropic and Google following OpenAI's Pentagon deal. This update primarily targets agentic knowledge work and computer-use tasks, allowing the model to interact via keyboard and mouse inputs based on screen captures. Key enhancements include the GPT-5.4 Thinking model's improved reasoning transparency and mid-process course correction, better context maintenance for long tasks, and an increased API context window of 1 million tokens. The model also features enhanced visual understanding, processing images up to 10.24 million pixels, and claims an 18 percent reduction in factual errors. These improvements aim to bolster OpenAI's competitive standing in capability, cost, and token efficiency.

Key takeaway

For Directors of AI/ML evaluating agentic model capabilities, GPT-5.4's explicit focus on computer-use tasks and its 1 million token context window warrant close examination. Your teams should assess its improved reasoning and visual understanding for long-horizon knowledge work automation, especially if current solutions struggle with context maintenance or factual accuracy. Consider piloting GPT-5.4 Pro via the API for enterprise applications requiring advanced agentic functionality.

Key insights

GPT-5.4 enhances agentic knowledge work through improved reasoning, context, visual understanding, and token efficiency.

Principles

• Agentic models benefit from transparent reasoning.
• Long-horizon tasks require robust context maintenance.

Method

GPT-5.4 operates by issuing keyboard/mouse inputs based on periodic desktop/application screenshots for computer-use tasks.

In practice

• Utilize GPT-5.4 Thinking for complex web research.
• Leverage 1M token context for extensive data analysis.

Topics

• GPT-5.4
• Large Language Models
• Agentic AI
• Token Efficiency
• Visual Understanding

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

Related on AIssential

• Introducing GPT-5.4 — GPT-5.4 enhances professional AI capabilities through integrated reasoning, coding, and agentic workflows with improved efficiency and safety.
• OpenAI launches GPT-5.4 with Pro and Thinking versions — GPT-5.4 offers enhanced capabilities, efficiency, and safety features for professional AI applications.
• GPT-5.4 Is INSANE! OpenAI's BEST Model Yet! 1 Million Context & Excels at Coding! (Fully Tested) — GPT 5.4 aims for a 1M context window and enhanced reasoning, excelling in complex code generation but with a distinct UI aesthetic.
• OpenAI launches GPT-5.4 Thinking and Pro combining coding, reasoning, and computer use in one model — GPT-5.4 unifies coding, reasoning, and computer operation, significantly boosting performance in professional knowledge work and desktop navigation.
• Everything We Know About OpenAI’s New GPT-5.4 Thinking Model — OpenAI's GPT-5.4 Thinking model enhances reasoning, coding, and agentic workflows for professional use.
• I Tried The New GPT 5.5 And I’m Never Going Back — GPT 5.5 excels at complex task execution and agentic workflows, moving beyond simple query responses.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI - Ars Technica.