Introducing our latest image generation model in the API

Source: OpenAI Date: 2025-04-23 URL: https://openai.com/index/image-generation-api

Summary

Summary

OpenAI released a new image generation model through the API in April 2025 — an updated DALL-E or next-generation image model with improved instruction following, style coherence, and text rendering within images. The release brought developer API access in line with the image generation capabilities that had been available in ChatGPT’s canvas and creative tools.

Implications

Platform/generative media thread. Image generation API parity with ChatGPT’s consumer tools matters for developer adoption: when ChatGPT has capabilities that aren’t yet in the API, developers working on image-generation applications are blocked or forced to work around limitations. The April 2025 timing coincides with strong developer demand for high-quality image generation following the improvements visible in GPT-4o’s image understanding and generation features. Competitive pressure from Midjourney, Stability AI, and Ideogram on image quality, and from Google Imagen and Flux on API availability, made this release a catch-up and parity move as much as a capability advance.