GPT-5.4 Nano (OpenAI)
GPT-5.4 Nano is OpenAI’s smallest GPT-5.4 tier for fast classification, extraction, and image understanding. Built for high-volume workflows.
Overview
GPT-5.4 Nano is OpenAI’s smallest GPT-5.4-class model. It generates text from your instructions and can analyze images you provide. It’s built for high-volume tasks like classification, data extraction, ranking, and lightweight coding support. It helps you ship fast automation without running a larger model on every step.
What you can build
- Screenshot Q&A for apps and dashboards
- OCR-style extraction of text fields from screenshots and photos
- Product and receipt tagging from images
- High-volume content classification and routing
- Fast entity extraction for CRM notes and support tickets
- Lightweight coding subagents for simple edits and checks
- Ranking and triage for queues, alerts, and moderation reviews
Inputs
- One or more image files for the model to read. Use PNG, JPEG, WEBP, or non-animated GIF. You can attach multiple images in one run.
- A required instruction describing what you want back. Ask for a summary, labels, a table, or a JSON object.
- Optional system-level instructions that set rules for tone and format. Keep this under 5,000 characters.
- Optional conversation identifiers if you want the model to keep chat history across turns. Provide a user identifier and a session identifier as a string or number.
- Optional reasoning effort setting that trades speed for deeper multi-step reasoning. Choose from none, low, medium, high, or xhigh.
- Optional verbosity control to keep answers short, balanced, or detailed.
- Optional web search toggle if you want the model to look things up online during the run.
Outputs
The model returns a text response. It can include structured sections like headings, bullets, and tables. If you ask for structured output, it can return machine-readable JSON for extraction and routing tasks. When you use chat history identifiers, the response also reflects prior turns in the same session.
Limitations
- Knowledge is not current past August 31, 2025. Use web search for newer facts.
- It does not support audio input or video input.
- Fine-tuning is not supported.
- Computer-use style UI control is not supported on this model.
- Image understanding depends on input quality. Blurry photos, tiny text, heavy compression, or skewed scans can reduce accuracy.
Safety & compliance
Follow OpenAI’s Usage Policies when you build with this model. Don’t use it for scams, impersonation, harassment, sexual exploitation of minors, weapons help, or other harmful activity. Add your own checks for high-stakes decisions, and keep a human in the loop when outcomes affect safety, jobs, housing, or health. For user-generated content, consider running text and images through OpenAI moderation before you store or act on results.
API quick start
Run gpt-5.4-nano with a single API call.
{
"prompt": "Tell me about Türkiye",
"inputImage": "https://your-cdn.com/input.png",
"user_id": "...",
"session_id": "..."
}curl -X POST "https://api.wiro.ai/v1/Run/openai/gpt-5-4-nano" \
-H "Content-Type: application/json" \
-H "x-api-key: YOUR_WIRO_API_KEY" \
--data-binary @- <<'JSON'
{
"prompt": "Tell me about Türkiye",
"inputImage": "https://your-cdn.com/input.png",
"user_id": "...",
"session_id": "..."
}
JSON