OpenAI has launched GPT-5.4 mini and nano models, expanding access to enhanced features for users across its platforms. This move follows the release of GPT-5.4 in early March.
The GPT-5.4 mini model is available for Free and Go users, enabling them to utilize capabilities that approach those of the full GPT-5.4 in certain areas. Users can access GPT-5.4 mini by selecting “Thinking” in ChatGPT’s menu. For paid subscribers, GPT-5.4 mini serves as a backup when their usage cap with GPT-5.4 is reached. OpenAI reports that GPT-5.4 mini outperforms the previous GPT-5.0 mini model in reasoning, multimodal understanding, and tool use.
Specifically, the improvements in GPT-5.4 mini include enhanced processing of non-text inputs like images and audio, as well as advanced web-search capabilities. The mini model operates more than twice as fast as its predecessor, making it a robust option for real-time applications.
The new GPT-5.4 nano model has also been introduced, optimized for tasks such as data classification and extraction. Unlike GPT-5.4 mini, the nano model is not integrated into the ChatGPT interface and is exclusively available through OpenAI’s API service. OpenAI has set the pricing for GPT-5.4 nano at $0.20 per million input tokens, indicating a focus on speed and cost-efficiency. The company anticipates that developers will delegate specific tasks to AI agents utilizing GPT-5.4 nano.
