OpenAI has massively upgraded ChatGPT’s image generation capabilities with GPT-4o, replacing the older DALL-E 3 model. This update delivers higher accuracy, better text rendering, enhanced photorealism, and improved object relationships in generated images. However, it may take slightly longer to produce these high quality results.
This advanced image-generation experience is rolling out to Plus, Pro, Team, and Free users, with Enterprise and Edu tiers gaining access soon. Developers will also get API integration in the coming weeks. To ensure safety and authenticity, OpenAI is embedding C2PA metadata into images, preventing misuse like deepfakes or objectionable content.
While a big leap forward, GPT-4o’s image generation still has limitations, such as struggles with cropping, precise edits, and rendering non-Latin text. Despite these, OpenAI is positioning ChatGPT as a strong rival to Gemini’s Flash 2.0 model.
