Here's how Canva's new image generation tool compares to models by OpenAI, Google, and Meta

Canva wants a attempt at winning the AI depiction generation race.

The scheme platform launched Canva Craft on Thursday, its archetypal image-generation model, akin to those from Google, Meta, and OpenAI.

Users can admission Craft through Canva’s Dream Lab, including alongside liberated accounts, although it is topic to AI use limits. You’ll need to sign in to a Canva document to get started.

In the publish announcement, Canva stated the tool plans a design’s layout, proportions, and content before generating an image, an method that aims to assistance it create legible lettering and exact details.

The business additionally stated that the example was trained on human-designed Canva templates, including posters, menus, and presentations, fairly than photographs.

To test how Canva performs, Business Insider created three depiction generation prompts alongside escalating difficulty levels — from designing a uncomplicated salutation cardstock to creating a photorealistic image from scratch.

Here is how Canva's brand-new depiction generation tool stacked up against three of the most powerful models.

Canva did not react to a petition for comment.

Starting off alongside an illustration

We started off uncomplicated alongside a immediate for a anniversary card.

Create a animation "Happy Birthday" salutation cardstock for a companion named Jessica who has pink hair, brown eyes, likes to deteriorate paisley, and is turning 28 years old in November.

Surprisingly, Canva, the design platform, cut abbreviated in its scheme of the anniversary card, from the clumsy font to the outsize caput of the animation character. It didn't equal annoy to contain any paisley.

By comparison, ChatGPT and Muse understood the assignment.

ChatGPT guessed that the recipient's favorite color power be pink, during Muse thoughtfully wove in cozy autumn elements for the November anniversary girl.

Meanwhile, Gemini took the petition exactly and produced an depiction of an genuine card. Extra points!

Winner: Muse

Editing an image

For the second task, I tested the AI models' capability to acknowledge an component and location it in a brand-new setting.

Edit this depiction of me holding a necklace pendant. Take the red garnet pendant off my hand and create it appearance as if it is displayed in a vitrine in a jewelry showroom.

Canva did a awesome job of identifying the pendant and creating an atmospheric setting, but, akin Gemini, it depicted the jewelry in a way that defied discipline and average sense. It would be unrealistic for a pendant to drift in the air or sit propped on a metal element pole.

ChatGPT and Muse formerly again outperformed another models in understanding how things typically appearance in reality. Muse, however, made several alterations to the pendant, which I didn't ask for.

Winner: ChatGPT

Creating a photorealistic depiction from scratch

With Halloween approaching, I asked for a realistic depiction of my favorite animal.

Generate a photorealistic depiction of a long-haired orange tabby cat in a Halloween costume alongside bat wings, perched on the border of a gothic medieval castle under the complete moon and a starless night sky.

Though the cat's fur looks a small too strandy and high-contrast in Canva's creation, no components are missing or malformed, and the depiction construction beautiful much matches the norm of Muse and Gemini.

The apparent victor for me in this round, however, would be ChatGPT.

Not lone is the cat's posture majestic, but the gentle glow on his remaining flank is extremely realistic, and the pale balance is fine done. There are additionally clouds quietly cradling the moon, which could be a reply to my petition for a starless sky.

Winner: ChatGPT

Overall, it's apparent that Canva motionless has a lot of area to develop compared to another models, particularly if explanatory scheme is its focus. For now, I'll rod to moving components about by hand.