Microsoft (MSFT, Financial)-backed OpenAI has upgraded its GPT-4o model with sharper image generation features as the race in AI chatbots heats up.
The updated system can now create more lifelike visuals, produce readable text in images, and allow step-by-step refinement through conversation. GPT-4o has replaced DALL·E 3 as the default image tool in ChatGPT and is now available to all users, including those on the free plan.
This aligns with OpenAI's broader aim to build one multimodal AI that can handle text, images, audio, and video. In the short term, it makes ChatGPT more useful. Long term, it supports OpenAI's goal of creating flexible AI systems for a range of tasks. A standout update is how closely the model now follows image prompts.
OpenAI said over 100 human trainers helped fine-tune it using feedback-based learning, improving things like faces, hands, and legible text. Businesses are already jumping in. GoDaddy (GDDY, Financial), for example, is using GPT-4o to make logos and stock visuals. Other uses include educational content and marketing design. OpenAI hopes this drives more adoption from teams and developers.
These tools also unlock revenue. OpenAI can sell API access, and offer branded models, and license-generated content. The update also reflects backing from Microsoft, which supports OpenAI's infrastructure through Azure. As AI tools advance, OpenAI is aiming to become a key part of how people create, work, and build.
