Try Pro FREE for 2 weeksGet Started
Start for free

gpt-image — Image Edit AI Model

GPT Image makes pictures from a prompt and edits existing ones from plain instructions — background removal, style transfer, object swaps, transparency where you want it. Billed to your own OpenAI account.

Try in Studio

What it does

GPT Image makes a picture from a written prompt, and edits one you supply on the optional image port - background removal, a style change, an object swapped out, transparency where you ask for it. Which of the two happens is decided by whether an image is wired, not by a mode setting.

It runs on OpenAI's servers against your own key, so there is no GPU to provision. Size, quality, output format and background are node settings, so what comes out is ready for the next step instead of needing a conversion pass.

When to use it

  • Product shots on a transparent background, ready to composite
  • Editing an existing image from a written instruction rather than by hand
  • Generating artwork for a page or a post from inside a flow
  • Cheap drafts first, then the same prompt again at a higher quality

Details

  • CategoryImage Edit
  • Year2026
  • LicenseOpenAI Commercial Terms

Compliance & Provenance

  • ProviderOpenAI · GPAI
  • EU AI Act RiskLimited Risk
  • Art. 50 TransparencyRequired — AI-generated outputs are marked

Inputs & Outputs

TextInputstring

Text prompt for image generation or editing instruction

ImageInputimageoptional

Input image for editing (up to 16 images, max 50MB each)

ImageOutputimage

Generated or edited image

Generated TextOutputstring

Text response from the model

Parameters

modelstringdefault GPT Image 2.5 Flareoptions: GPT Image 2.5 Flare, GPT Image 2.5 Sunburst, GPT Image 2.0, +3 more

Model variant

Which GPT Image model to call. 'GPT Image 2.5 Flare' (default) — the 2.5 line's fast variant: better images than 2.0 at roughly half the latency, same per-token rate; the right pick for everyday generation. 'GPT Image 2.5 Sunburst' — the 2.5 line's quality variant: tighter control over edits and subject preservation across multiple turns, at longer generation times; pick it when editing precision matters more than speed. '2.0' / '1.5' / '1' — earlier generations, kept for flows already pinned to them. '1 Mini' — cheapest and fastest, suitable for thumbnails or drafts. The 'xhigh' and 'max' quality levels require a 2.5 model.

sizestringdefault autooptions: auto, 1024x1024, 1536x1024, 1024x1536

Output size

Output image dimensions. 'auto' (default) lets the model choose based on prompt. '1024x1024' square, '1536x1024' landscape, '1024x1536' portrait.

backgroundstringdefault autooptions: auto, transparent, opaqueadvanced

Background transparency

Background handling. 'auto' (default) lets the model decide. 'transparent' produces a transparent background — requires output_format png or webp. 'opaque' forces solid background.

qualitystringdefault mediumoptions: max, xhigh, high, medium, low, autoadvanced

Quality (max/xhigh/high/medium/low/auto — max and xhigh need a 2.5 model)

Generation quality vs cost/speed. 'medium' (default) = balanced. 'low' = fastest and cheapest, suitable for drafts. 'high' = more detail, slower and costlier. 'xhigh' and 'max' push detail further still and are only accepted by 'GPT Image 2.5 Flare' and 'GPT Image 2.5 Sunburst' — sending them to an older model variant makes the OpenAI call fail. 'auto' lets the model pick, which is OpenAI's own default. Quality drives output token count, and output tokens are what this node is billed on, so 'max' costs several times a 'low' run.

output_formatstringdefault jpegoptions: jpeg, png, webpadvanced

Output format

Encoded file format. 'jpeg' (default) for smallest files (no transparency). 'png' for lossless with transparency support. 'webp' for modern lossy/lossless with transparency. Pick png/webp when background='transparent'.

moderationstringdefault autooptions: auto, lowadvanced

Content moderation

Strictness of OpenAI's content filter. 'auto' (default) is standard. 'low' is less restrictive — only choose when working with adult or borderline imagery within OpenAI's policy.

Limitations

  • The xhigh and max quality levels are accepted only by the 2.5 tiers; sending one to an older tier fails the call.
  • A transparent background needs png or webp - jpeg has nowhere to write the transparency.
  • Runs on your own OpenAI API key. Without one configured in the workspace the node cannot run.

Tags

  • image-generation
  • image-editing
  • multimodal
  • text-to-image

Alternatives in Image/Video Models

  • Nano Banana

    Google Gemini-based image gen/edit with multimodal reasoning. Needs API key.

  • Omni

    Google Gemini Omni: text/image/video in, video out, with conversational editing. Which input ports apply depends on Task — Text to Video: Text only. Image to Video: + Image 1. Reference to Video: + Image 1–3. Edit / Extend: + Video (aspect ratio and resolution are ignored). Unused ports are dropped even if wired. Needs API key.

  • VEO

    Text/image to video with native audio, now run by Gemini Omni 1.1 Flash (Google retires VEO 3.1 on 2026-10-22). Needs API key.

  • Video Analysis

    Gemini video analysis. Feed a video and get back timecoded highlight segments. Needs API key.

Resources

All AI models