Introducing GPT-5.4 mini and nano
OpenAI has just announced the release of GPT-5.4 mini and nano, representing a significant step towards more accessible and efficient large language models. These models are designed to bring the capabilities of the powerful GPT-5.4 to applications where speed and cost are paramount.
Key Improvements & Performance
GPT-5.4 mini boasts significant improvements over its predecessor, GPT-5 mini, across several key areas: coding, reasoning, multimodal understanding, and tool use. Notably, it achieves speeds over 2x faster than GPT-5 mini. Impressively, it approaches the performance of the larger GPT-5.4 model on benchmarks like SWE-Bench Pro and OSWorld-Verified.
GPT-5.4 nano, the smallest and most cost-effective version, is a substantial upgrade over GPT-5 nano. It's recommended for tasks like classification, data extraction, ranking, and for powering coding sub-agents handling simpler tasks.
Here's a performance comparison (based on OpenAI's data):
| GPT-5.4 (xhigh) | GPT-5.4 mini (xhigh) | GPT-5.4 nano (xhigh) | GPT-5 mini (high¹) | |
|---|---|---|---|---|
| SWE-Bench Pro (Public) | 57.7% | 54.4% | 52.4% | 45.7% |
| Terminal-Bench 2.0 | 75.1% | 60.0% | 46.3% | 38.2% |
| Toolathlon | 54.6% | 42.9% | 35.5% | 26.9% |
| GPQA Diamond | 93.0% | 88.0% | 82.8% | 81.6% |
| OSWorld-Verified | 75.0% | 72.1% | 39.0% | 42.0% |
1 The highest reasoning_effort available for GPT-5 mini is 'high'.
Use Cases
These models are particularly well-suited for scenarios where low latency is critical. Think coding assistants demanding responsive feedback, sub-agents swiftly completing supporting tasks, systems processing screenshots, and multimodal applications requiring real-time image reasoning.
OpenAI highlighted strong early feedback from customers like Hebbia, CodeRabbit, Mercor, GitHub, Rox, Notion and Whoop. Aabhas Sharma, CTO at Hebbia, noted that GPT-5.4 mini delivered “strong end-to-end performance… matched or exceeded competitive models on several output tasks and citation recall at a much lower cost.”
Coding Focus
GPT-5.4 mini and nano excel in coding workflows, enabling fast iteration on targeted edits, codebase navigation, front-end generation, and debugging. This makes them a strong choice for coding tasks needing speed and cost efficiency. The models consistently outperform GPT-5 mini at similar latencies and are rapidly approaching GPT-5.4's coding capabilities.
Availability & Pricing
Details regarding availability and pricing were not extensively covered in the initial announcement. Further information is expected to be released by OpenAI in the coming weeks. Developers should keep an eye on the OpenAI API documentation for updates.
This release represents a strategic move by OpenAI, recognizing that the 'best' model isn't always the largest, but rather the one that balances performance with speed and cost-effectiveness.