I compared Bard and Midjourney AI images using similar prompts.
The world of AI image generators has been very busy recently. This week, tech giants Google and Midjourney both released major updates to their AI image models. Google released its much-anticipated feature on Bard that can generate AI images. At the same time, Midjourney launched the next iteration of its popular AI art generator, V6.
The question now is: which AI tool creates better images?
In this article, I will compare the image results generated by Bard and Midjourney using similar text prompts.
Let’s get started.
Prompt: wine pours from a bottle into a glass, mockup, photo, minimalism, banner, plain background
Both AIs did a good job; the pictures look kind of the same. But, I like the light and small details in Midjourney’s picture more. The bubbles look so real.
Prompt: create Valentine’s Day themed sticker sheet, cartoon, desserts, hearts, teddy bears, flowers, chocolate, strawberry, heart shaped balloons, sticker outline, on light pink background
These stickers are so cute! They’re perfect for Valentine’s Day. Which one’s better? I think more people might like the left picture because it has more details. But if you like simple, cartoony stuff, you’ll like the right image.
Prompt: a painting of a white dog standing on the beach by the sunset in the background, in the style of colorful complexity, celestial, heavy impasto, the stars art group (xing xing), brushwork mastery, heavy impasto brushwork, swirling vortexes
Both did an excellent representation of the text prompt. But one weird thing I noticed in both images is a fake signature in the right corner. That’s strange. Artist’s signatures appearing on AI images are still an unsolved issue.
Prompt: an isometric 3d model of a man sat in the entrance of a camping tent, sitting in a mini world, popout effect in 3d
In this example, the man in the left image is not sitting but rather lying. So there’s a clear winner here.
Prompt: an image of a simple and minimalistic logo of a startup called “Gen AI”
I like Bard’s logo more. It’s simpler and looks better for a product.
Prompt: gangster fat orc sitting on a sofa, cyberpunk, shadowrun fat orc sitting on sofa, fat gangster orc
Even though Bard’s result is more vibrant, I still prefer the quality of Midjourney’s. The orc on the right looks more bad-ass.
Prompt: An image of a cute anthropomorphic silver-white kitten with round pupils, wearing a bright yellow turtleneck sweater, a white jacket and jeans outside, sitting on an airplane seat. The seat is near the window, with a passenger sitting next to it. The kitten looks out the window.
I was surprised to see how close the results were. But I personally prefer the result from Midjourney because of the better shadows and finer details.
Prompt: an image of a modern bathroom in the style of Oscar Niemeyer, either a very porous pale white travertine wall and wooden ceiling completely open on au luxuriant tropical Brazilian garden
Okay, I think there’s no competition here. The lighting, the details, and the overall aesthetics on the right image look so photorealistic.
Prompt: crochet doll of Vincent Van Gogh holding a paint brush in art studio
I’m on the fence with this one. The right image’s vibrancy is eye-catching, but the left’s subdued colors and simplicity give it a more authentic feel.
Prompt: Motion blurred light trails a man is walking in the streets of chennai wearing a coat and a hat black and white image motion blurred camera shake images
This one’s a matter of personal taste. For me, the right image, with its distinctive style, stands out.
If you are not familiar with Bard yet, it is a free AI chatbot from Google. You can access Bard here, log in with your Google account, and start chatting with the AI or generating images.
Midjourney on the other hand, is a freemium AI image generator that is accessible either on Discord or on the Midjourney website. It costs 10 USD per month.
I can’t show you the image generation support on the Midjourney website just yet because I don’t have access to it at the moment.
After comparing AI-generated images across 10 prompt categories, Midjourney’s V6 edges out Google’s Bard as the superior tool for AI images. Midjourney generated more photorealistic and creative images overall.
However, Bard should not be underestimated as it has strong potential given that it was just unveiled. With further development, it could give Midjourney a run for its money.
Which one I’d recommend?
Both tools have pros and cons. Bard is free and easy to use while Midjourney produces better results but requires a paid subscription. Your choice comes down to budget, quality needs, and personal preference. The true winner is us users who get to leverage these amazing AI tools for creative projects.
If you are interested on the comparison of Bard and Dall-E 3, let me know in the comments.
Software engineer, writer, solopreneur