Canva released Craft on Thursday, its first in-house image generation model and, in Canva's words, the first foundation model built and trained from scratch by Canva Research. Craft debuted as the fourth most advanced image generation model on DesignArena's leaderboard, behind three models from OpenAI, which Canva says makes its research group the second most advanced AI lab for image generation.
Craft places higher in individual categories. Canva says it ranks first for People & Portraits out of 87 models, first for Landscapes & Nature out of 85 models, second for Abstract & Patterns and fourth for Products. The three models ahead of it overall are GPT-Image-2.5 Sunburst, GPT-Image-2.5 Flare and GPT Image 2.
Craft generates images in two stages. The first interprets the prompt and plans the layout, including object placement, relative scale, visual hierarchy and any text the image needs to carry. The second stage generates the image itself.
Key facts
- 01Canva released Craft on Oct. 22, 2026, the first foundation model built and trained from scratch by Canva Research.
- 02Craft debuted fourth on DesignArena's image generation leaderboard, behind three OpenAI models.
- 03It ranks first on DesignArena for People & Portraits (out of 87 models) and Landscapes & Nature (out of 85 models).
- 04Craft outputs up to 2K resolution across more than 30 aspect ratios and renders text in 55 languages.
- 05Cheaper Craft-fast and Craft-ultrafast variants and Canva MCP Server support are listed as coming soon.
Canva built Craft for design work rather than photography. The company trained it on millions of human-made templates from its content library, including posters, social posts, presentations, infographics, menus and typography-heavy layouts. Professional designers assessed more than 25,000 pairs of images in head-to-head comparisons against leading models during development.
Output runs up to 2K resolution across more than 30 aspect ratios, and users can enter custom ratios. Prompts can be written in any language, and Craft can render text inside images in 55 languages. Users can edit specific elements without changing the rest of the image, and results can be converted into layered, editable designs with Canva's Magic Layers.
Canva says Craft is available to all Canva users globally through Dream Lab, subject to account and usage limits, and to developers through Canva's Public API. Support through the Canva MCP Server, which would let agents in chatbots such as ChatGPT and Gemini use the model, is listed as coming soon.
Two cheaper variants, Craft-fast and Craft-ultrafast, are also in development. Canva says they will deliver high-quality image generation at a fraction of the cost of leading frontier models.
Researchers from Leonardo.ai, which Canva acquired in 2024, contributed to the model. A Canva spokesperson said building and owning these models gives the company complete control over cost, quality and speed, and the in-house model could lower what Canva pays to serve image generation through its API.
Canva says it has more than a quarter of a billion users, who create over 400 designs every second. The company reached about $4 billion in revenue last year, though it has cut its 2026 growth forecast, and outside analysts have reportedly lowered their estimate of its valuation from $42 billion to $35 billion.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
