© ROOT-NATION.com - Use of content is permitted with a backlink.
OpenAI has officially closed the gap in its model lineup by introducing GPT-5.4 Mini and GPT-5.4 Nano. These solutions are focused on speed and efficiency when working with large volumes of tasks. The new models are designed to compete with high-performance systems like Gemini 3 Flash.
Back in August 2025, along with the launch of GPT-5, the company also showed the compact GPT-5 Mini and GPT-5 Nano models, designed specifically for handling large-scale workloads. Since then, OpenAI has introduced several flagship updates, including GPT-5.1, GPT-5.3, and GPT-5.4, but no new small models have been released during this period. This pause allowed competitors, including Google, to strengthen their position in the segment with efficient compact solutions such as Gemini 3 Flash.
Now OpenAI has joined the race and presented GPT-5.4 Mini and GPT-5.4 Nano. It positions them as the company’s most powerful small models to date, specially optimized for intensive workloads. In particular, the GPT-5.4 Mini demonstrates significantly better results in programming, logical thinking, multimodal understanding, and tool handling than the GPT-5 Mini, while being more than twice as fast. In a number of popular artificial intelligence benchmarks, including SWE-Bench Pro and OSWorld-Verified, this model almost reaches the level of the much larger GPT-5.4.
Read also: Trump vs. Claude AI: How the AI Debate Is Unfolding in the U.S.
The GPT-5.4 Nano model is the most affordable in the GPT-5.4 series and also significantly outperforms the previous GPT-5 Nano in terms of performance. OpenAI notes that it is best suited for typical tasks, including classification, data extraction, ranking, and the work of auxiliary agents in programming.
Read also: Camera, biometrics, and ChatGPT: What we know about the first OpenAI device
The new GPT-5.4 Mini is already available in ChatGPT, Codex, and via the developer API. Just like the older GPT-5.4 model, it supports text and image input, working with tools, calling functions, web search, file search, computer interaction, and additional features. The context window is 400 thousand tokens, and the cost is $0.75 for 1 million incoming tokens and $4.50 for 1 million outgoing tokens. For comparison, GPT-5 Mini cost $0.25 for 1 million incoming tokens and $2 for 1 million outgoing tokens.
For developers using Codex, CLI, IDE extensions, or the web version, GPT-5.4 Mini consumes only 30% of the GPT-5.4 quota, making it convenient for tasks that don’t require complex thinking. Ordinary users can use this model in ChatGPT in the Free and Go plans through the Thinking feature. For GPT-5.4 subscription holders, the Mini is a backup model for GPT-5.4 Thinking.
At the same time, OpenAI emphasized that GPT-5.4 Nano will not be available in ChatGPT or Codex. This model is offered exclusively through the API at a price of $0.20 for 1 million incoming tokens and $1.25 for 1 million outgoing tokens.
Read also:
- Best Free AI Tattoo Design Generators: Which Produce Ink-Ready vs Screen-Only Art?
- Claude developer Anthropic sues US government over “supply chain risk” designation