openai

OpenAI: GPT-5.4 Nano

The OpenAI GPT-5.4 Nano is a large language model that can process text, images, and files as input. It has a context length of 400,000 tokens and supports tools and reasoning capabilities. This model may be worth considering for those who prioritize robust processing capabilities and are willing to pay $1.25 per million output tokens. Its benchmark blended score is 35.3 across one independent metric, suggesting it performs well in this area; however, more comprehensive testing would provide a clearer picture of its overall strengths and weaknesses.

Quality Score
100/100
price + capability + benchmarks
Input Price
$0.20
per 1M tokens · checked 2026-09-23
Output Price
$1.25
per 1M tokens · checked 2026-09-23
Context Window
400,000
tokens

Benchmark results

Independent, published benchmarks. Blended score 31.5 across 1 benchmark, last refreshed 2026-09-23. How scoring works →

BenchmarkMeasuresScore
AI Index broad capability composite 34.2
Model ID
openai/gpt-5.4-nano
Vendor
openai
Released
March 2026
Tokenizer
GPT
Input Modalities
file, image, text
Output Modalities
text
Max Output
128,000 tokens
Tool Calling
✓ supported
Structured Output
✓ supported
Reasoning Mode
✓ supported
Vision
✓ accepts images
Audio
no
Moderated
yes

What it costs in practice

Computed from the current $0.20/M input and $1.25/M output rates. Run your own numbers →

JobTokensCost
Summarize a 50-page report 30k in / 1.5k out under $0.01
Classify 1,000 customer emails 500k in / 50k out $0.16
A month of a busy support chatbot 5M in / 2M out $3.50

Similar models

Quick answers

How much does OpenAI: GPT-5.4 Nano cost?
$0.20 per million input tokens and $1.25 per million output tokens.
What is OpenAI: GPT-5.4 Nano's context window?
400,000 tokens, roughly 600 pages of text in a single request.
Does OpenAI: GPT-5.4 Nano support tool calling?
It supports tool calling, structured output, a reasoning mode.
Can OpenAI: GPT-5.4 Nano process images?
Yes, it accepts image input.

The Model Movers Report

One email every Friday, built from this site's own rankings: the current top five by benchmark score, every model released in the last seven days, and one note worked out from that week's numbers. You can unsubscribe from any issue with one click.