OpenAI released GPT-5.4, its latest frontier model featuring specialized Pro and Thinking versions for enterprise use. The model features a context window of one million tokens. This capacity more than doubles its predecessor and matches capabilities from rivals Google and Anthropic.
The new Thinking mode utilizes extreme reasoning to solve multi-step scientific and research problems. This feature allocates additional computation and time to process complex tasks. The expanded context window allows the model to analyze significantly larger documents and data sets.
OpenAI adopted a strategy of frequent, iterative updates to manage market expectations. The rollout responds to increasing competition from Anthropic Claude models.