Tasks and inputs
Canva Craft is listed for Text-to-image. It accepts text and produces image. Use the publisher’s input format and preprocessing for your chosen deployment.
Canva Craft is Canva's first in-house image generation model: a planner plus renderer that outputs up to 2K images with legible text in 55 languages.
The specification score summarizes recorded features and limits. It is not a benchmark result.
Pricing conditions: Indicative: about 1 cent per image in AI Credits in Dream Lab; API pricing not published. Confirm on your own Canva account.
Recorded modality support shows what this model can accept and produce. An undocumented modality is shown as unknown rather than assumed to be unsupported.
Accepts
Produces
Generated images usually fail designers in the same place: the words. A poster headline comes back as near-letters, and someone rebuilds it by hand. Canva Craft, released on 22 October 2026 and built by Canva Research, is designed around that failure. It is Canva's first image-generation foundation model of its own.
Craft works in two stages. A planning model writes a composition plan, deciding what goes where and which words must appear, and a rendering model draws the picture from that plan. Splitting the job is why Canva can claim legible text in 55 languages rather than a handful.
That design suits layout-heavy work: flyers, social tiles, menus, event banners. It is less obviously an advantage for painterly art with no text in it.
| Item | Detail |
|---|---|
| Developer | Canva Research |
| Released | 22 October 2026 |
| Input / output | Text prompt in, image out |
| Maximum resolution | Up to 2K |
| Aspect ratios | 30+ presets plus custom ratios (in Dream Lab) |
| Text rendering | 55 languages |
| Where | Dream Lab for all Canva users; Canva's public API |
| Indicative cost | About 1 cent per image in AI Credits (Dream Lab) |
| Weights | Closed |
Canva says Craft debuted fourth on a public image leaderboard. Faster Craft-fast and Craft-ultrafast variants are planned, and an MCP server for AI assistants is listed as coming soon.
Developer details are thinner than consumer ones. Reports circulate an API model name of craft_v1_2k and an API limited to six aspect ratios, possibly in preview. Neither appeared on a Canva page we read, so treat both as unconfirmed. Before you design around the 30-plus ratios Dream Lab offers, call the API and see which ones it actually accepts. API pricing was not published at launch; check your developer account.
In Dream Lab, Craft draws more of the monthly AI allowance than simpler tools. If a team generates heavily, that allowance, not the per-image cent, may be the limit you hit first.
Text is the thing to test, because it is the claim.
If Craft gets the words right more often, the time saved in correction is likely worth more than any difference in per-image price. The full field is under text-to-image models.
Compare the model on the work you need to complete.
Canva Craft is listed for Text-to-image. It accepts text and produces image. Use the publisher’s input format and preprocessing for your chosen deployment.
Choose the output size and quality your design needs. Include drafts and revisions when budgeting an accepted image.
Use the model identifier and account requirements for the selected API route. Confirm its request limits and output format with one complete task before connecting a larger workflow.
These records share a directory category. Check the source card, output type and deployment requirements before treating a model as a replacement.
recraft
Recraft V4.1 Flash V2 is Recraft's fastest, cheapest image model: 1K raster output in about 1.1 s for $0.007, quality close to Recraft V4.1.
MiniMax
MiniMax H3 Image is an open-weight 33.1B image generation and editing model producing native ~16MP output in six steps, with up to 16 reference images.
MiniMax
MiniMax H3 Image Highspeed is the faster hosted tier of MiniMax's H3 Image generator and editor, at double the image-token price of the standard tier.
OpenAI
GPT-Image-2.6 Flare is OpenAI's fast image generation and editing model for interactive iteration, with SVG, animated GIF and markup-guided edits.
OpenAI
GPT-Image-2.6 Sunburst is OpenAI's detail-first image generation and editing model, adding SVG, animated GIF output and markup-guided edits.
krea
Krea 3 is Krea's in-house image model for text-to-image and instruction-based edits at 1K or native 2K, priced from $0.03 per image.