OpenAI has unveiled its latest offerings, the GPT-5.4 Mini and Nano models, accessible to free users starting today. This announcement comes shortly after the launch of the more advanced GPT-5.4 model, with the Mini version designed as a ‘smaller, cheaper’ alternative for applications where speed and cost efficiency are critical.
The GPT-5.4 Mini boasts enhanced capabilities across various tasks, including coding, reasoning, multimodal understanding, and tool utilization. OpenAI claims that this new model operates over two times faster than its predecessor, approaching performance levels similar to the flagship GPT-5.4 on specific benchmarks such as SWE-Bench Pro and OSWorld-Verified.
In contrast, the GPT-5.4 Nano is characterized as the smallest and least expensive variant of the GPT-5.4 series, intended for tasks where rapid performance and cost are paramount. The company notes that the Nano model is particularly useful for functions like classification, data extraction, ranking, and creating coding subagents.
“These models are built for the kinds of workloads where latency directly shapes the product experience: coding assistants that need to feel responsive, subagents that quickly complete supporting tasks, computer-using systems that capture and interpret screenshots, and multimodal applications that can reason over images in real-time,” OpenAI stated in a recent blog post. This emphasis on speed highlights the company’s commitment to improving user experience in various applications.
The GPT-5.4 Mini is also being positioned as a ‘subagent’ within systems like Codex. Developers are encouraged to leverage this model for tasks involving planning, coordination, and final judgment, while delegating narrower, parallel subtasks—such as searching a codebase or reviewing extensive documents—to the more responsive GPT-5.4 Mini.
Notably, the Mini version demonstrates a strong ability to perform complex computer-related tasks, including the rapid interpretation of screenshots from intricate user interfaces. This capability enables it to reason over images in real-time, effectively narrowing the performance gap with the more powerful GPT-5.4 on visual benchmarks like OSWorld-Verified.
Access to the GPT-5.4 Mini has been granted to all ChatGPT free and Go users through the ‘Thinking’ feature located in the ‘+’ menu. Meanwhile, the Nano model will be exclusively accessible via the OpenAI API.
This strategic rollout of the Mini and Nano models underscores OpenAI’s ongoing efforts to diversify its product offerings and meet the evolving needs of developers and users alike. As AI technologies continue to advance, the introduction of these models may further empower businesses and individuals to integrate AI into their workflows, enhancing productivity and efficiency.
See also
OpenAI Unveils GPT-5.4 Mini and Nano for Faster, Cost-Effective AI Workflows
OpenAI Launches GPT-5.4 Mini and Nano Models for High-Volume AI Tasks at Lower Costs
Scaling GenAI: Achieving 5,000+ Concurrent Video Generations with Asynchronous Architecture
Hardware-Optimized Diffusion Model IP Market Surges to $2.34B in 2025 with 25.8% CAGR
OpenAI Reveals DALL-E 3 with Enhanced Prompt Accuracy and 4K Output Capabilities
