← All models

GPT-5.4 nano

Published Publisher OpenAIgpt-5.4-nano

Overview

GPT-5.4 nano is the smallest and least expensive GPT-5.4 variant, designed for simple, high-volume tasks where speed and cost matter. OpenAI recommends it for classification, data extraction, ranking, and coding subagents handling simpler supporting work. (openai.com)

Capabilities and limits

  • Accepts text and image inputs and produces text output. Audio and video are not supported.
  • Supports configurable reasoning effort, streaming, function calling, structured outputs, and tools including web search and file search.
  • Context window: 400,000 tokens; maximum output: 128,000 tokens. Knowledge cutoff: August 31, 2025.
  • Fine-tuning, computer use, and tool search are not supported. Availability and rate limits depend on the API usage tier. (developers.openai.com)

Release and availability

Released March 17, 2026. The launch announcement described GPT-5.4 nano as API-only. The dated snapshot is gpt-5.4-nano-2026-03-17. (openai.com)

Lifecycle status

As of November 16, 2026, OpenAI marks both the model ID and dated snapshot as deprecated. Shutdown is scheduled for April 1, 2027; OpenAI lists gpt-6-luna as the replacement. (developers.openai.com)

Pricing

OpenAI API text-token pricing, checked November 16, 2026: USD $0.20 per 1 million input tokens, $0.02 per 1 million cached input tokens, and $1.25 per 1 million output tokens. The cached-input rate applies to eligible cached tokens; the input and output figures are the standard rates, not Batch API rates. Regional processing/data-residency endpoints carry a 10% uplift. Tool use may incur additional charges. (developers.openai.com)

Official sources