In this OpenAI Build Hour, presenters describe GPT-5.4, which they say supports up to a million tokenThe chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.Full definition → of context windowThe maximum amount of text a model can consider at once — its working memory for the current conversation or task.Full definition →, native computer use and a new tool search feature. OpenAI says advanced builders were giving agents hundreds of tools, which caused context management problems, meaning the model's working text had to hold descriptions of every tool.
Tool search changes that by hiding the descriptions and letting the AI agentAn AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.Full definition → discover the ones it needs. A presenter calls the idea progressive disclosure: not every tool needs to be in context for every task. This is the company's account of the design and its motivation. The excerpt does not show measurements of how much it helps.
Terms in this piece · Glossary
token — The chunk of text a model reads and writes in — roughly three-quarters of a word — and the unit AI usage is billed in.
context window — The maximum amount of text a model can consider at once — its working memory for the current conversation or task.
AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
agent harness — The scaffolding around a model that turns it into a working agent — the loop, the tools it can call, and the rules for when to stop.