Canva releases its first image generation model
techcrunch ·
Canva has launched its first proprietary image generation model, named Canva Craft, designed to compete with offerings from major tech firms like Google and OpenAI. The model operates in two stages: first interpreting user prompts to plan spatial relationships and text, then generating images up to 2K resolution in various aspect ratios. By training on human-designed templates rather than photographs, the system aims to produce accurate details and legible text that can be easily edited or converted into layered designs. Canva Craft is available to all users via the Dream Lab tab and will also be accessible through the API and third-party chatbots. The development benefits from research contributions by Leonardo.ai, which Canva acquired in 2024, helping the company reduce API costs while improving quality control. Additionally, Canva plans to release faster, cost-efficient versions of the model called Craftfast and Craft-ultrafast soon. This launch occurs as the company adjusts its growth forecast for 2026, leading to a reported drop in valuation from $42 billion to $35 billion.
techcrunch