OpenAI: GPT-5.4 Mini Boosts ChatGPT with Speed

OpenAI continues to innovate with the launch of the GPT-5.4 mini and GPT-5.4 nano models, aimed at enhancing the efficiency of ChatGPT. These new models are designed to deliver improved performance, focusing on speed and efficiency.

The GPT-5.4 mini model is now available on ChatGPT for both free users and those on paid versions. This model, while smaller than GPT-5.4, stands out for its speed, being more than twice as fast as its predecessor, GPT-5 mini. It excels in areas such as coding, reasoning, multimodal understanding, and tool usage. OpenAI emphasizes that for certain applications, response speed is more critical than model size. On some benchmarks, GPT-5.4 mini even approaches the performance of GPT-5.4, the larger model.

To use GPT-5.4 mini on ChatGPT, users can access the "Thinking" feature from the + menu on the free version and ChatGPT Go. In other versions, it serves as a fallback solution when the GPT-5.4 Thinking model reaches its usage limits. Additionally, GPT-5.4 mini is offered on OpenAI's API within the Codex tool, allowing for broader integration for developers.

Meanwhile, OpenAI is introducing GPT-5.4 nano, an even more compact model, available exclusively through its API. This model is designed for developers seeking a fast and cost-effective AI for specific tasks such as classification, data extraction, and managing simple agents.

These launches come amid increased competition in the AI market, with players like Claude gaining popularity among businesses, and Gemini experiencing exceptional growth in 2025. OpenAI is thus strengthening its position by offering diverse and high-performing solutions tailored to the varied needs of users and developers.