October 25, 2026
The digital design platform “Canva” announced the launch of its first model for generating images with artificial intelligence under the name “Canva Craft,” in a move aimed at enhancing its competitiveness against major global technology companies, such as “Google,” “Open AI,” and “Meta.”
According to what was reported by the “TechCrunch” website, the company’s internal research team developed this new model, focusing on improving the quality of images, the accuracy of details, and providing the ability to modify specific elements within the design in isolation from the rest of the components.
The model’s work depends on two basic stages: The first begins with carefully analyzing text commands, determining the distribution of elements, their sizes, and spatial relationships between them, then moving to the stage of generating the final image with a resolution of up to 2K, while providing support for about 30 different image aspect ratios.
The company indicated that “Canva Craft” trains on designs actually completed by humans, including posters, presentations, publications, and information graphics, going beyond the exclusive reliance on photographs. This approach gives the model a greater ability to deal with written texts inside images, which is one of the most prominent challenges facing visual content generation tools currently.
The new model is now available to all platform users through the “Dream Lab” section, including those with free accounts within the specified usage limits, in addition to making it available to developers through a special programming interface.
In the same context, the company intends to integrate the capabilities of the new model into external services that include artificial intelligence assistance, in addition to working to develop faster and less expensive versions during the next stage.
