Canva Research · Released 2026-10-22
Canva Research’s image foundation model plans composition before rendering up to 2K images. The generation task returns an image and optional asset; Magic Layers conversion into an editable design is a separate task and credit use. Publisher language/rank claims are not our Chinese-text tests.
Original model explanation ↗No comparable leaderboard score is included yet. Explore the evaluations and examples below.
Published official-source analysis; APIs not called
The result types make the distinction concrete: generation supplies an image; a successful image-to-design job supplies a Canva design and editing link.
Preview API docs, not called
Not a claim of native Craft layer output
Publisher illustration, not our test or native Craft layer output
A balloon, a background and a heading can share one image.
Publisher examples without per-image generation logs
Originally announced October 22. Craft generation and later Magic Layers conversion have separate interfaces, access conditions and credits; our October 26 analysis is not a second model launch.
Original release ↗