Design platform Canva has released its first image generation model, Canva Craft, a system its research team built to compete with models from Google, OpenAI, ByteDance and Meta. The model does something most image generators don't: it plans the design before it draws anything.
The process happens in two stages. In the first, the model interprets a user's prompt and produces a plan that covers the spatial relationships between elements, their relative scale, the visual hierarchy and the text that will appear in the image. In the second, it generates the final image at up to 2K resolution, in more than 30 aspect ratios meant for different platforms.
Canva says that structure helps the model get complex images right, especially those that need readable text and objects placed exactly where they belong. Users should also be able to make targeted edits without changing the rest of the picture.
The company trained the model on designs rather than photographs. Canva says it used “millions of human-made templates” from its content library, including social media posts, presentations, infographics, menus, posters and typography-heavy layouts. Images created with Canva Craft can be turned into layered, editable designs using Canva Magic, the company's image-to-design tool.
The launch builds on a steady push into AI. In April, Canva added prompt-based editing, which lets users create or change a design by describing the change. The platform also introduced layers for AI-generated images, so users can alter one part of a picture while leaving the rest alone. Last year, Canva released its own model that understands design structure and layers.
On Design Arena, a crowdsourced leaderboard that ranks AI-generated designs, Canva Craft currently sits fourth. Three spots above it are OpenAI models: GPT-Image-2.5 Sunburst, GPT-Image-2.5 Flare and GPT-Image-2.
Canva says the model includes “robust input and output moderation, privacy and safety controls, and reporting tools.” The company also plans to let AI agents connect to Canva Craft through the open Model Context Protocol, and it intends to support image generation inside third-party chatbots such as ChatGPT and Gemini through the same protocol.
Cost is a factor too. Canva has relied on several image models across its platform, and the new model could help reduce the company's API costs. Canva said the model also gives it better control over quality and speed. Researchers from Leonardo.ai, which Canva acquired in 2024, contributed to the work.
Canva is also developing two lower-cost versions, called Craft-fast and Craft-ultrafast. The company says they will be released soon.
The launch comes at a difficult moment for the company's finances. Canva was on track for $4 billion in revenue last year. It has, however, cut its growth forecast for 2026, and external analysts reportedly estimate its valuation has fallen from $42 billion to $35 billion.
Canva Craft is available now to all users globally. To try it, open the three-dot menu on the left-hand side of Canva and select Dream Lab.
Dieser Artikel wurde mit Unterstützung von KI verfasst.
News Factory APP - agentische News für besseres SEO & AEO.