OpenAI’s March 5, 2026 announcement presents GPT-5.4 as its new flagship model for professional work, saying it combines stronger reasoning, coding, tool use, and agent-style workflows into one system that is meant to handle documents, spreadsheets, presentations, and longer, more complex tasks with less back and forth. It says GPT-5.4 Thinking in ChatGPT can now show an upfront plan while working, while the API and Codex versions add native computer-use abilities, a 1 million token context window in some settings, and improved tool search so agents can find and use the right tools more efficiently. The article also highlights better benchmark results than GPT-5.2 in areas like knowledge work, coding, browsing, and computer use, and says the model is more factual and more token-efficient overall. OpenAI says GPT-5.4 is rolling out across ChatGPT, the API, and Codex, with GPT-5.4 Thinking replacing GPT-5.2 Thinking for Plus, Team, and Pro users, while GPT-5.4 Pro is offered for people who want maximum performance on harder tasks. It also notes that API pricing is higher than GPT-5.2, though OpenAI argues the model’s efficiency can reduce total token use on many jobs

Recent news