Doha: OpenAI has announced a new update to its ChatGPT artificial intelligence (AI) robot, allowing users to generate images based on detailed, complex, and unconventional instructions, marking a significant development in the company’s flagship product.

According to Qatar News Agency, the new version of ChatGPT is based on the GPT-4o model, which enables the robot to analyze text and images together within a single, integrated system, giving it an unprecedented ability to generate more complex and accurate images. The company highlighted that users can now, for example, describe a four-panel cartoon, specifying the characters appearing in each panel and what they say, and ChatGPT will instantly create a complete cartoon based on these details.

While previous versions of ChatGPT were capable of generating images, they did not have the ability to accurately and reliably integrate multiple and diverse concepts into a single image, nor did they have the ability to handle text within images as accurately as the new update. The GPT-4o model enables a more interactive image editing experience, allowing users to request an image to be created or uploaded, then provide subsequent instructions to modify it, such as changing colors or adding new details, without having to rewrite the entire description.

Gabriel Goh, a researcher at OpenAI, explained that this technology represents a completely new type of AI, adding that the company’s models no longer separate text generation from image generation, but rather combine the two processes together to produce smoother and more accurate results. Goh noted that traditional image generation models have always struggled to generate images of unfamiliar concepts, such as a triangular-wheeled bicycle, but the new version of ChatGPT is now able to handle these complex requests with ease.

OpenAI announced that the new update to ChatGPT will be available to all users, whether through the free version or through paid subscriptions.