OpenAI debuts GPT-5.4 mini and nano for speed and affordability
OpenAI just rolled out two new AI models, the GPT-5.4 mini and nano, on March 18, 2026.
Both are built for speed and affordability, making them great for things like coding assistants or real-time image tasks where quick responses matter.
The mini model is almost as powerful as the full-sized GPT-5.4 but runs more than twice as fast as the old GPT-5 mini.
GPT-5.4 mini outperforms predecessor nano cheaper
The GPT-5.4 mini outperforms the previous version in key benchmarks, so it's a strong pick for coding or reasoning-heavy work.
If you just need something simple (like sorting data), the nano model keeps things even cheaper and lighter.
Pricing is student-budget-friendly: $0.75 per million input tokens ($4.50 per million output tokens) with the mini, while nano goes for just $0.20 per million input tokens ($1.25 per million output tokens), though you'll need API access to use it, making advanced AI more accessible than ever without breaking the bank!