GPT Mini is a small-footprint variant of the GPT large-language-model family aimed at lower-cost, lower-latency use cases. Community benchmarks suggest it lags behind larger contemporaries such as Claude and Gemini 3 Flash on multi-step agent tasks.