OpenAI has launched its latest models, GPT-5.4 mini and nano, designed for developers prioritizing faster responses and lower costs. Available now, GPT-5.4 mini operates over twice as fast as its predecessor and maintains competitive performance on key benchmarks, scoring 54.4% on SWE-Bench Pro. The nano version excels in efficiency for simpler tasks like data extraction, with competitive scores while significantly reducing costs. GPT-5.4 mini is priced at $0.75 per million input tokens and $4.50 for outputs, while nano costs just $0.20 and $1.25. Both support text and image inputs along with a 400,000 token context window. Developers can effectively utilize a multi-model workflow, balancing larger models for complex tasks and smaller ones for execution. This approach has proven beneficial, with early feedback indicating that GPT-5.4 mini often outperforms competitors at a lower cost. Both models are accessible via API, Codex, and ChatGPT, offering new opportunities for real-time AI applications.
Source link
Share
Read more