Video & Creative AI
Canva launches Craft with pre-generation planning
The Verge AIhigh confidence
Why it matters
- →Pre-generation planning addresses the core failure mode of current image models: lack of structural control.
- →Training on curated templates rather than raw web data improves text legibility and design coherence.
- →Global availability makes this a direct competitor to Midjourney and DALL-E for professional design workflows.
Canva has launched its first proprietary AI image generation model, Canva Craft, which utilizes a two-stage architecture to plan designs before generating images. Trained on millions of human-made templates from its platform, the model aims for greater precision and consistency, supporting up to 2K resolution and over 30 aspect ratios. It currently ranks fourth on Design Arena’s leaderboard, behind OpenAI’s GPT-Image-2.5 variants. The tool includes robust safety controls and is available globally via Canva’s Dream Lab feature.
Read originalThe story around this
Earlier coverage that leads up to this article, and what followed. Lines connect each piece to the closest one after it, converging here.
