OpenAI launches GPT 5.4 Mini, Nano: Faster, Cheaper AI models for real-time tasks

OpenAI has unveiled two new compact versions of its flagship AI models.

The newly released versions, referred to as “GPT 5.4 mini and GPT 5.4 nano,” are aimed at delivering near-flagship performance at a minimum cost with significantly faster speeds.

Released on March 17, the newer models are customized for developers and businesses that need high-volume, low-latency AI operations like real-time coding assistants, data extraction, and agentic workflows, without the expense of running large models.

Performance and speed

OpenAI cited that GPT-5.4 mini operates twice as fast as its predecessor, GPT-5 mini, while closely matching the capabilities of the full-sized GPT-5.4 model. On the Graduate-Level Google-Proof Q&A (GPQA) Diamond reasoning test, the model scored 88% in contrast to the full model, which achieved 93%.

The even smaller GPT-5.4 nano is designed for automating simpler and repetitive tasks such as classification and data ranking, providing the lowest cost and fastest response times.

Use cases and availability

OpenAI stated that “workloads where latency directly shapes the product experience.” This includes “vibe coding” assistants that function in real time and computer-use agents that manage data entry.

GPT-5.4 mini is accessible via the API, the Codex coding tool, and ChatGPT. Through the “thinking feature”, free and go-tier users can try it.

Pricing

GPT-5.4 mini is available at $0.75 per million input tokens and $4.50 per million output tokens. GPT-5.4 is even cheaper, priced at $0.20 for input and $1.25 for output. If compared, the full ChatGPT 5.4 model is valued at around $2.50 and $15.00, respectively.