OpenAI’s Next AI Model Will Have ‘Extreme’ Reasoning

OpenAI’s next GPT model is coming—and soon, according to a person with knowledge of it.

Among the highlights, the new model, GPT-5.4, will have more than double the context window of the current GPT-5.2 model. That means the model can handle queries with many more words or data, up to 1 million tokens compared to 400,000 tokens today, the person with knowledge said. (As a reminder, a token is a word or part of a word.) That’ll put it on par with rival models from Google and Anthropic that already support a 1 million token context window. (Past OpenAI models like GPT-4.1 have also supported a 1 million token context window but 5.2 doesn’t.)

The new model is also better at tasks that can take as long as a few hours, meaning that it’s more capable of remembering information like the details of its user’s request and what it’s allowed and not allowed to do across many steps, and is less prone to mistakes, they said. That could prove helpful for applications like OpenAI’s Codex coding tool, which use AI to automate very complex, long-running tasks.

Another notable update is the addition of an “extreme” reasoning mode, allowing the model to spend significantly more time (and therefore, compute) on tough questions, the person said.