OpenAI Updates ChatGPT's Image Generator

OpenAI has rolled out an updated version of the image generator built into ChatGPT. Users can now upload a sketch or reference photo, or simply describe an idea in words, and the tool produces a more personalized, polished image that stays closer to the original intent.

Earlier versions of the model often filled in details on their own, so the final image could drift from what the user actually wanted. This update shifts that balance: results now follow the uploaded materials — a sketch, a photo, a reference — more closely instead of replacing them with the model's own take. That matters for practical work: fewer retries, fewer edits, a finished image faster.

Business owners in Kazakhstan should look at where a tool like this AI agent for business already fits — social media mockups, product cards, banner drafts, or visualizing an idea for the team. At 1Cifer we plug tools like this into an employee's AI agent as a skill: a marketing agent, for instance, doesn't need to switch between services — the right skill for the task connects itself.