OpenAI launches GPT-5.4 Mini and Nano to power faster high-volume AI workloads

OpenAI has introduced GPT-5.4 Mini and GPT-5.4 Nano, two new compact models aimed at speed, scale, and efficiency. Positioned as lighter counterparts to its flagship GPT-5.4 model, both versions are designed for environments where response time and cost matter as much as capability.

Not every AI task needs the full weight of a top-tier model. Many workflows, especially in coding, automation, and real-time systems, benefit more from quicker turnaround than from maximum depth. GPT-5.4 Mini and Nano are built to slot into those gaps without feeling like a compromise.

GPT-5.4 Mini and Nano: Built for speed

GPT-5.4 Mini is the more capable of the two, offering a noticeable step up from GPT-5 Mini across coding, reasoning, multimodal understanding, and tool use. OpenAI says it runs more than twice as fast while approaching the performance of the full GPT-5.4 model in benchmarks like SWE-Bench Pro and OSWorld-Verified.

That translates to smoother coding workflows, faster debugging loops, and more responsive AI assistants. It is also designed to work as part of a broader system. Larger models can handle planning and decision-making, while the mini model executes smaller, clearly defined tasks such as scanning files, navigating codebases, or processing supporting data.

GPT-5.4 Nano, on the other hand, strips things back even further. It is built for lightweight, high-frequency tasks like classification, data extraction, ranking, and basic coding support. Think of it less as a generalist and more as a background worker handling repetitive jobs at scale.

ALSO READ: OpenAI brings Codex to Windows with native agent sandbox and parallel AI agents

Both models also lean into multimodal use cases. They can interpret dense user interface screenshots and support computer-based workflows, which is increasingly relevant as AI tools move beyond text and into everyday digital environments.

Price and availability

OpenAI is clearly pitching these models as cost-efficient options for developers and businesses running large-scale operations.

GPT-5.4 Mini is available across API, Codex, and ChatGPT, priced at $0.75 (approx. Rs 69) per million input tokens and $4.50 (approx. Rs 415) per million output tokens. GPT-5.4 nano is API-only, with pricing set at $0.20 (approx. Rs 18) per million input tokens and $1.25 (approx. Rs 116) per million output tokens.

ALSO READ: You can now use Shazam in ChatGPT; here’s how

There are a few practical touches worth noting. In Codex, GPT-5.4 mini uses just 30 per cent of the GPT-5.4 quota, making it easier to delegate simpler tasks without burning through usage limits.

In ChatGPT, it appears as an option under the “Thinking” menu for Free and Go users, and also acts as a fallback when higher-tier models hit rate limits.

With these two models, OpenAI is filling out the middle and lower tiers of its model stack, giving developers more flexibility to match the right model to the right job without overpaying for it. What are your thoughts? Drop a comment below.