Microsoft Copilot gets ChatGPT-like viral image generator

OpenAI’s image generator went viral back in March when ChatGPT switched to GPT-4o. No separate DALL·E model, no janky renderings—just smart, context-aware image creation baked right into the chat. It worked. Over 700 million images were made in a week. People loved it.

Fast-forward to one and nearly half months, Microsoft is bringing a ChatGPT-like image generator to Copilot. The news came via the official Copilot X handle, which confirmed that Copilot can now:

• Render readable, accurate text in images
• Follow layered, complex prompts
• Edit what you’ve already created
• Transform styles using references or tweaks

Basically, all the fun stuff ChatGPT users have had for weeks is now available in Copilot. And to be fair, it works well. The new image generation in Copilot feels snappier, more precise, and finally useful for more than just random art prompts.

You can edit styles, rework parts of an image, and actually use the output in something meaningful. That said, Microsoft’s pace still feels a step behind.

At its 50th anniversary event, the company shared several updates around Copilot, but most of them sounded like reactions to what OpenAI and Google already launched months ago.