Canva Craft: Canva's first in-house image model plans before it renders — up to 2K in more than 30 aspect ratios, millions of its own templates in the training data, live for all users

“Canva built its first image model from scratch with millions of its own human-made templates in the training data — and since October 22 it's live for every Canva user in Dream Lab.”

Usable claims

• On Thursday, October 22, 2026 (calendar-verified; the date carried by TechCrunch's byline and The Verge's 'according to an announcement on Thursday' — the Canva page itself carries no date label), Canva published 'Introducing Canva Craft'. Per the page: 'Today, we're unveiling Canva Craft, our first state-of-the-art image generation model, developed and trained by Canva Research, our in-house AI lab.' The architecture, per the page: 'In the first stage, we designed an image planning model to interpret the prompt and write a detailed plan of the composition: which elements appear, their spatial relationships, relative scale, visual hierarchy, and any text that needs to appear in the image.' and 'In the second stage, a dedicated image rendering model translates this plan into a high-fidelity image at up to 2K resolution across more than 30 aspect ratios.' Training, per the page: 'Its training data includes millions of human-made templates from Canva's content library, spanning professionally designed posters, social posts, presentations, infographics, menus and typography-rich layouts.' plus designer feedback: 'Canva's professional designers worked alongside dozens of researchers to assess more than 25,000 pairs of images in head-to-head comparisons with leading models.' Languages: 'Prompts can be written in any language, and Craft can create text in 55 languages.' The page's own DesignArena readings, all of one crowdsourced leaderboard: 'It debuted as the fourth most advanced image generation model on DesignArena, delivering industry-leading visual quality and accurate text rendering to support a wide range of creative tasks.'; 'Craft outputs rank #1 spot globally for both People & Portraits (out of 87 models) and Landscapes & Nature (out of 85 models).'; 'It also ranks #2 for Abstract & Patterns and #4 for Products, placing it among the world's best models across multiple subcategories.'; and the lab claim 'Built specifically for design, Craft makes Canva the world's second most advanced AI lab for image generation, as measured by rankings from DesignArena, the largest crowdsourced assessor of AI design.' Speed, per the page: distillation cut the inference steps 'to a fraction of the original model', and a custom inference engine makes Craft an order of magnitude faster than Canva's first internal version — the baseline is Canva's own earlier version, never a competitor. Safety: 'Like all AI products on Canva, Craft is backed by Canva Shield, our comprehensive approach to AI safety.' Availability: 'Canva Craft is available starting today to all Canva users globally.' with Dream Lab as the entry point; 'Craft is also available via Public API and the Canva MCP Server…' in one sentence, while the page's feature line separately labels 'Canva MCP (Coming Soon)' — the card carries the MCP Server as coming-soon per the page's own label; and 'These advances underpin Craft-fast and Craft-ultrafast, two upcoming variants designed to deliver high-quality image generation at a fraction of the cost of leading frontier models.' — no ship date stated. Per TechCrunch (Ivan Mehta, 12:00 PM PDT October 22): 'The company noted that rather than using photographs, it used designs to train this model using human-designed templates from Canva's library, including posters, social posts, presentations, infographics, menus, and typography-rich layouts.' — 'rather than using photographs' is TechCrunch's rendering where the Canva page's verb is 'includes'; the Leonardo layer ('In 2024, Canva acquired media generation platform Leonardo.ai. It said that the researchers from Leonardo contributed to the new image model.') is TechCrunch's and absent from the page; and TechCrunch renders the variants as 'Craftfast and Craft-ultrafast, which will be released soon'. Per The Verge (Emma Roth, October 22): 'The model is currently ranked fourth on Design Arena's leaderboard for the best AI image generation models, a crowdsourced benchmark for AI designs. It sits behind OpenAI's GPT-Image-2.5 Sunburst, GPT-Image-2.5 Flare, and GPT-Image-2.' — The Verge's own leaderboard reading at its publish time.