← All models

GPT-4o image generation

Published Publisher OpenAI

Model Overview

GPT-4o image generation is a multimodal image generation capability natively embedded in GPT-4o that can generate or transform images based on text and image inputs. OpenAI released this capability on March 25, 2025. (openai.com)

Capabilities and Uses

  • Generate photorealistic or other styles of images from natural language and render text within images.
  • Modify images over multiple turns using conversation context, with uploaded images as editing subjects or visual references.
  • Follow instructions involving multiple objects and constraints, suitable for tasks such as infographics, logos, diagrams, comics, character design, and creative compositions. (openai.com)

Limitations and Safety

The release announcement noted that the model may still produce improper cropping, hallucinations, object attribute binding errors, difficulties with precise drawing, multilingual text rendering errors, and insufficient editing precision or handling of dense small-text information. Complex image generation may take longer; at release, the announcement stated that rendering often took up to about a minute. Inputs and outputs are subject to OpenAI safety policies and moderation mechanisms. (openai.com)

Release and Availability

At its release on March 25, 2025, the feature began rolling out to ChatGPT Free, Plus, Pro, and Team users and was also available in Sora; support for enterprise and education users was still forthcoming at the time. On April 23, 2025, OpenAI announced that the native multimodal model powering the ChatGPT image experience at the time was available through the API as gpt-image-1. That API model was shut down on October 23, 2026; this does not change the model_id: null in this record's input. (openai.com)

Pricing

Historical API prices, in US dollars (USD): OpenAI's official model page for gpt-image-1 lists text input at $5 per million tokens and cached text input at $1.25 per million tokens; image input at $10 per million tokens and cached image input at $2.50 per million tokens; and image output at $40 per million tokens. The page does not list a price for text output. These are prices for the shut-down gpt-image-1 API model; they do not represent prices currently available for purchase and are not ChatGPT subscription prices. (developers.openai.com)

Pricing source: OpenAI GPT Image 1 model page; access date: November 19, 2026.

Official Sources