OpenAI is broadening access to its latest AI capabilities by releasing GPT-5.4 mini, a streamlined version of its advanced GPT-5.4 model, now available to ChatGPT Free and Go users. This move delivers enhanced reasoning, multimodal understanding, and faster performance to users who previously had limited options, narrowing the gap between free and paid tiers.
GPT-5.4 mini can be selected via the ”Thinking” option in ChatGPT’s plus menu and serves as a fallback for paid users who hit their GPT-5.4 usage limits. Compared to the earlier GPT-5.0 mini, the new version processes images and audio inputs with more sophistication and performs quicker web searches. OpenAI claims that GPT-5.4 mini runs more than twice as fast as GPT-5.0 mini while offering improvements in complex reasoning and tool usage, which are essential for professional tasks like programming and data analysis.
GPT-5.4 nano offers high-speed API access for developers
Alongside GPT-5.4 mini, OpenAI has introduced GPT-5.4 nano, optimized for high-speed, cost-effective tasks such as data classification and extraction. However, nano is not integrated into the ChatGPT chatbot experience; instead, it is available exclusively through OpenAI’s API, focusing on empowering developers to deploy AI agents efficiently.
Priced at $0.20 per million input tokens, the GPT-5.4 nano model aims to minimize operational costs in AI workflows while maintaining respectable performance. OpenAI anticipates that developers will harness GPT-5.4 nano to offload simpler or repetitive processes, freeing up more powerful models for demanding use cases.
By introducing these tiered releases, OpenAI is making its high-end AI increasingly accessible without sacrificing speed or capability, responding to the growing demand for smarter AI in everyday and professional contexts.