Canva Steps Up Its AI Ambitions with the Launch of Canva Craft to Rival Industry Heavyweights

Popular visual communication platform Canva is diving even deeper into the artificial intelligence landscape with the official introduction of Canva Craft, a proprietary image generation model designed to compete directly with advanced offerings from tech giants like OpenAI, Microsoft, and Google. Announced on Thursday, the new model marks a significant milestone for the Australian-headquartered design company as it transitions from integrating third-party generative technologies to deploying its own sophisticated, purpose-built architecture tailored specifically for creators, marketers, and everyday users.

At the core of Canva Craft is an innovative two-stage architecture that fundamentally alters how text-to-image systems operate. Unlike traditional models that attempt to render an entire image in a single pass, Canva Craft is engineered to plan a comprehensive design blueprint before generating the final visual output. According to the company’s official announcement, this structured approach allows for significantly greater precision, consistency, and control over the final asset. The model’s two-stage system is particularly beneficial for complex images that demand legible text rendering and precise object placement—areas where conventional generative AI tools have historically struggled.

Capable of producing high-definition visuals at up to 2K resolution across more than 30 distinct aspect ratios, Canva Craft is designed to handle a vast array of creative tasks. Beyond standard social media posts and digital marketing content, the model is built to generate realistic scenes, intricate digital art, expansive environmental imagery, and much more. This versatility reflects Canva’s broader mission to serve as an all-in-one platform where individuals and enterprise teams can ideate, design, and execute projects without needing to switch between disparate software applications.

The debut of Canva Craft builds upon a sustained and aggressive push into artificial intelligence by Canva over the past year. In April, the platform rolled out advanced prompt-based editing capabilities that enable users to create assets or make targeted modifications to existing designs simply by describing the desired changes in plain language. Furthermore, the design tool previously introduced a breakthrough "Magic Layers" feature in beta, which successfully brought traditional graphic design layers to AI-generated images. This functionality allows users to isolate and manipulate specific elements of an AI-generated graphic while leaving the surrounding background or composition completely untouched, bridging the gap between raw generative outputs and practical, professional-grade design workflows.

What distinguishes Canva Craft from many competing models is the unique dataset used to train it. Canva revealed that the system was trained on millions of human-made templates sourced directly from the platform’s extensive content library. This vast repository encompasses a wide variety of professional design formats, including social media layouts, corporate presentations, detailed infographics, restaurant menus, and marketing collateral. By grounding the AI in actual human design principles and structured templates, Canva has equipped the model with an inherent understanding of visual hierarchy, composition, and professional aesthetics.

Early indicators suggest that Canva Craft can hold its own against established market leaders. On Design Arena’s leaderboard—a prominent crowdsourced benchmark platform used to evaluate the efficacy and quality of AI image generation models—Canva Craft currently holds the fourth position globally. It sits closely behind top-tier competitors, trailing only OpenAI’s GPT-Image-2.5 Sunburst, GPT-Image-2.5 Flare, and GPT-Image-2. As crowdsourced evaluations continue to test the limits of generative AI, Canva’s strong initial ranking demonstrates the viability of its specialized, template-trained approach.

Alongside creative capabilities, Canva has emphasized safety, governance, and interoperability in the rollout of its new model. The company noted that Canva Craft comes equipped with robust input and output moderation mechanisms, advanced privacy and safety controls, and comprehensive reporting tools designed to mitigate misuse and ensure responsible deployment across global markets. Looking ahead, Canva has outlined plans to expand the ecosystem surrounding the model by allowing external AI agents to plug directly into Canva Craft through the open Model Context Protocol. Additionally, the company intends to introduce faster iterations of the technology, including more efficient Craft-fast and Craft-ultrafast models, to cater to high-volume commercial workflows.

Canva Craft is available now to all users globally, marking its immediate integration into the platform’s daily operations. Creators and professionals looking to test the new model can access it by navigating to the three-dot option located on Canva’s left-hand menu and selecting Dream Lab.