GPT-5.4 mini is OpenAI’s lower-latency, lower-cost variant of the GPT-5.4 family, available via the OpenAI API. It trades some capability for significantly faster responses and cheaper per-token pricing, making it well suited to high-volume applications, chat interfaces, and real-time use cases.
Developers typically use mini for production workloads where cost-per-request matters and the task complexity doesn’t require the full flagship model.