GPT-5.4 nano
OpenAI · United States · 2026
The cheapest model of its generation — 20 cents per million tokens in, for classification and other high-volume work.
GPT-5.4 nano was released on 17 March 2026 together with GPT-5.4 mini, and sits at the bottom of the generation’s price ladder: $0.20 per million input tokens, $0.02 cached, and $1.25 per million output. OpenAI positions it for work where speed and cost matter more than depth — classification, extraction, routing and other high-volume tasks. The specification is otherwise the same as the mini variant: a 400,000-token context window, 128,000 tokens of output, text and image input, a knowledge cutoff of 31 August 2025, and support for reasoning tokens. That last point is worth noting — the cheapest tier of the family still reasons, which was not true of the nano models in the GPT-4.1 generation. Relative to base GPT-5.4 it is roughly twelve times cheaper on both input and output, which places the whole GPT-5.4 line across a 150-fold price range from nano to Pro.
▸Videos
No videos yet.
▸Reviews
No reviews yet. Be the first!