OpenAI Launches GPT-5.4 Mini and Nano for Faster Coding

OpenAI has released two new compact models in the GPT-5.4 family: GPT-5.4 mini and GPT-5.4 nano. Both are built for high-volume use cases where low latency matters most, including coding assistants, multimodal apps, and agentic workflows.

GPT-5.4 mini sits between the flagship GPT-5.4 and the older GPT-5 mini. OpenAI says it delivers improvements in coding, reasoning, multimodal understanding, and tool use, running at more than twice the speed of its predecessor. On benchmarks like SWE-Bench Pro and OSWorld-Verified, the mini model approaches full GPT-5.4 performance.

GPT-5.4 nano is the smallest and lowest-cost model in the series. OpenAI recommends it for classification, data extraction, ranking, and simple coding subtasks within larger agent systems.

OpenAI highlighted the pattern in its Codex platform, where a hierarchy of models works in tandem: GPT-5.4 handles planning and final judgment, while mini and nano take on parallel subtasks like codebase search, large file review, and document processing.

This reflects a broader trend in AI product development, where teams combine models of different sizes to balance performance and cost. The pattern gives developers more flexibility to scale AI-assisted coding without paying flagship prices for every single step.