OpenAI Rolls Out ChatGPT Images and GPT Image 1.5 for Faster, More Accurate Image Generation

December 16, 2025

Details

  • OpenAI introduces ChatGPT Images, featuring a streamlined image creation and editing workflow built on the new GPT Image 1.5 model.
  • The model delivers image render times four times faster than the previous DALL·E 3, while preserving lighting, composition, and subject consistency across prompts and edits.
  • Improved instruction adherence allows users to make detailed changes, such as altering the background to dusk, without inadvertently affecting other image areas.
  • An updated 'Images' tab launches on both web and mobile versions of ChatGPT, enabled through a simple app update.
  • API access via the gpt-image-1.5 endpoint gives developers the same tools for image generation and iterative editing available in ChatGPT's chat interface.
  • OpenAI claims the new model excels at diverse applications, from product mock-ups to photorealistic portraits and stylized art, expanding its usefulness beyond traditional digital artwork.
  • The rollout begins December 16, 2025, for all ChatGPT users, both free and paid, with enterprise rate limits aligned to existing GPT-4o token quotas.
  • Safety features from DALL·E 3, including watermarking and content filtering for violent or illicit imagery, remain in place to address concerns around deepfakes and misuse.

Impact

This launch intensifies competition with image AI platforms like Midjourney and Stability AI, which still require manual editing for precise changes. OpenAI’s move to integrate image tools directly in ChatGPT encourages multimodal use, which may drive longer engagement and increase subscriber retention. The advance also positions OpenAI as a proactive leader on content safety, aligning with growing regulatory scrutiny on synthetic media provenance and transparency.

Read more