OpenAI today announced the release of GPT‑5.4 mini and GPT‑5.4 nano, its most capable small models to date. The mini model improves on the earlier GPT‑5 mini across coding, reasoning, multimodal understanding and tool use, running more than twice as fast while approaching GPT‑5.4 performance on benchmarks such as SWE‑Bench Pro and OSWorld‑Verified. The nano version is the smallest, cheapest option, priced at $0.20 per 1M input tokens and $1.25 per 1M output tokens, and is recommended for classification, data extraction, ranking and simple coding sub‑tasks. Both models are available in the API, Codex and ChatGPT; the mini supports text and image inputs, tool use, web search, file search, computer use and a 400k context window, costing $0.75 per 1M input tokens and $4.50 per 1M output tokens. The mini is also offered in Codex, CLI, IDE extension and web, using only 30% of the GPT‑5.4 quota and enabling developers to delegate narrow subtasks to a cheaper model. In ChatGPT, the mini is available to Free and Go users via the “Thinking” feature and as a rate‑limit fallback for other users. The nano is only available in the API. OpenAI notes that these models are designed for latency‑sensitive workloads such as coding assistants, sub‑agents, computer‑using systems and real‑time multimodal applications, where speed and cost are critical. The company also references a System Card addendum on its Deployment Safety Hub for safeguards information.
Introducing GPT-5.4 mini and nano
OpenAI has released GPT‑5.4 mini and GPT‑5.4 nano, the company’s smallest and fastest models yet. The mini model improves on the earlier GPT‑5 mini in coding, reasoning, multimodal understanding and tool use, running more than twice as fast while approaching GPT‑5.4 performance on benchmarks such as SWE‑Bench Pro and OSWorld‑Verified. The nano version is the smallest, cheapest option, aimed at classification, data extraction, ranking and simple coding sub‑tasks. Both models are available in the API, Codex and ChatGPT, with the mini supporting text and image inputs, tool use, web search, file search, computer use and a 400k context window. Prices are $0.75 per 1M input tokens and $4.50 per 1M output tokens for the mini, and $0.20/$1.25 for the nano.
Why it matters
OpenAI’s new small models enable faster, lower‑cost coding assistants and sub‑agents, improving latency‑sensitive workflows and allowing developers to compose systems that use larger models for planning and smaller ones for execution.