GPT-4o powered image generation by OpenAI in ChatGPT: goodbye, DALL-E | OpenAI Chat | ChatGPT download | Chat GPT login free | Turtles AI

GPT-4o powered image generation by OpenAI in ChatGPT: goodbye, DALL-E
The GPT-4o model introduces more detailed and precise images, accessible to all users with new editing and visual creation capabilities
Isabella V25 March 2025

 

OpenAI has enhanced ChatGPT by integrating image generation using the GPT-4o model, which is now available to all users. This advanced technology improves the accuracy and quality of the images created, opening up new possibilities for both consumers and businesses.

Key Points:

  • Integration of GPT-4o into ChatGPT: OpenAI has implemented the GPT-4o model into ChatGPT, enabling image generation directly within the chatbot interface.
  • Availability to all users: The feature is available to users on Free, Plus, Team, and Pro plans, expanding access to this advanced technology.
  • Image quality improvements: GPT-4o provides more accurate rendering of images, with significant improvements in the consistency of detail and the ability to generate readable text within images.
  • Safeguards against misuse: OpenAI has implemented measures to prevent misuse of the technology, including inappropriate content blocks and adding standard metadata to generated images to indicate their AI origin.

OpenAI broadcast a short live broadcast on March 25, 2025, announcing the integration of the GPT-4o model into ChatGPT, bringing advanced image generation capabilities directly into the chatbot interface. OpenAI certainly covered a lot of ground in about 15 minutes. Sam Altman and team debuted native image generation in the 4o model. They then showed off some demos, and before it was over, we were testing the feature in the ChatGPT iPhone app. This innovation, called “Images in ChatGPT,” is now available to all users, regardless of subscription plan, including Free, Plus, Team, and Pro tiers. Described as an “omnimodal” model, GPT-4o can generate a variety of data types, including text, images, audio, and video. One of the key improvements is the model’s ability to maintain accurate relationships between attributes and objects in the images it generates. For example, GPT-4o can accurately handle complex requests involving multiple elements, such as “a blue star and a red triangle,” avoiding confusion common in previous models. In a demo, the OpenAI team took a selfie and then asked ChatGPT to transform it into an “anime style.” It took several seconds, but it did indeed generate what was requested. Additionally, the model showed significant improvements in rendering text within images, enabling the creation of visual content with legible and consistent text. The image generation process uses an autoregressive approach, creating images sequentially from left to right and top to bottom, similar to writing text. This method differs from traditional diffusion models and helps improve accuracy in rendering text and consistency of visual elements. While this approach may take slightly longer to generate, OpenAI believes that the higher quality of the images produced more than makes up for the additional time. To ensure responsible use of the technology, OpenAI has implemented robust security measures. The system is designed to prevent watermark removal, block inappropriate content generation, and reject requests to create images that do not comply with the guidelines. Additionally, all generated images include standard C2PA metadata to indicate that they were created by OpenAI AI. Users retain ownership of the generated images and are free to use them in accordance with OpenAI’s usage policies.

This evolution represents a significant step in integrating multimodal capabilities into chatbots, opening up new opportunities for both consumers and businesses in the creation of visual content.

Video