OpenAI launches GPT 5.4 mini and nano models for high-volume workloads: Pricing, availability and more

OpenAI has launched two new AI models — GPT-5.4 mini and GPT-5.4 nano — aimed at delivering faster performance and lower costs for high-volume workloads. In its official announcement, the company said that the new models bring several capabilities of its larger GPT-5.4 system into smaller, more efficient versions designed for tasks where speed and scale are important. OpenAI said GPT-5.4 mini improves on its earlier mini model across coding, reasoning, and image understanding, while running more than twice as fast. The model is also claimed to perform close to the larger GPT-5.4 system in some benchmark tests.GPT-5.4 nano, the smallest version, is designed for tasks such as classification, data extraction, and simple coding operations. The company said it is suited for use cases where low cost and quick responses are required.The new models are designed for applications where response time affects user experience. These include coding assistants, automation tools, and systems that analyse images or screenshots.OpenAI said GPT-5.4 mini performs well in coding workflows such as editing code, debugging, and navigating codebases. The model is also used in setups where larger models handle planning tasks while smaller models manage specific subtasks.The company said GPT-5.4 mini supports multimodal inputs, including text and images. It can interpret screenshots and assist in computer-based tasks.In internal tests, the model showed improved performance over earlier versions in tasks related to system interaction and interface understanding.GPT-5.4 mini is available in OpenAI’s API, Codex, and ChatGPT. The model supports features such as tool use, web search, and file handling, and comes with a 400,000 token context window.The pricing is set at $0.75 per million input tokens and $4.50 per million output tokens.GPT-5.4 nano is available through the API and is priced lower at $0.20 per million input tokens and $1.25 per million output tokens.