AI news
December 27, 2023

Google Imagen 2 VS OpenAI Dall-E 3 - Battle of AI Image Generators

The battle between two of the best AI image generators. Who does it better?

Jim Clyde Monge
by 
Jim Clyde Monge

Google has recently announced the next version of its own AI image generator, Imagen. Imagen 2.0 makes crazy realistic images — I can’t even tell what’s real or fake anymore.

In this article, I wanna put Imagen 2 up against the other big AI image generator Dall-E 3 by OpenAI and really see what they both can do.

In the comparison test, I will be using the example images released by Google on its Imagen 2 announcement and ChatGPT for Dall-E 3. I will add all the image prompts so you can try them yourself.

Let’s get started.

Prompt #1

A shot of a 32-year-old female, up and coming conservationist in a jungle; athletic with short, curly hair and a warm smile

A shot of a 32-year-old female, up and coming conservationist in a jungle; athletic with short, curly hair and a warm smile

Well, it’s pretty obvious. The image from Imagen 2 looks more realistic than the one from Dall-E 3, especially in regards to skin texture and hair details, which AI image generators often struggle with.

Prompt #2

Small canvas oil painting of an orange on a chopping board. Light is passing through orange segments, casting an orange light across part of the chopping board. There is a blue and white cloth in the background. Caustics, bounce light, expressive brush strokes

Small canvas oil painting of an orange on a chopping board. Light is passing through orange segments, casting an orange light across part of the chopping board. There is a blue and white cloth in the background. Caustics, bounce light, expressive brush strokes

The soft tones on the left image give it a photorealism effect. But Dall-E 3 made a more accurate representation of the blue and white cloth in the background.

Prompt #3

The robin flew from his swinging spray of ivy on to the top of the wall and he opened his beak and sang a loud, lovely trill, merely to show off. Nothing in the world is quite as adorably lovely as a robin when he shows off — and they are nearly always doing it.

The robin flew from his swinging spray of ivy on to the top of the wall and he opened his beak and sang a loud, lovely trill, merely to show off. Nothing in the world is quite as adorably lovely as a robin when he shows off — and they are nearly always doing it.

Imagen 2 delivered an incredibly life-like robin in this prompt, comparable to a National Geographic photograph. However, Dall-E 3’s interpretation feels more whimsical and expressive, capturing the lovely essence described. Different approaches, both impressive

Prompt #4

A cup of strawberry yogurt with the word “Delicous” written on the side, sitting on a wooden tabletop. Next to the cup of yogurt is a plate with toast and a glass of orange juice.

A cup of strawberry yogurt with the word “Delicous” written on the side, sitting on a wooden tabletop. Next to the cup of yogurt is a plate with toast and a glass of orange juice.

This example reveals inconsistencies in Dall-E 3’s text generating capabilities, as it failed to include the “Delicious” detail on the yogurt cup specified in the prompt. Imagen 2 performed accurately here.

Prompt #5

An abstract logo representing intelligence for an enterprise AI platform, “Vertex AI” written under the logo.

An abstract logo representing intelligence for an enterprise AI platform, “Vertex AI” written under the logo.

Okay, this is really cool. Both AI systems impressively rendered the “Vertex AI” text. However, Imagen 2 edges ahead with its minimalist, professional logo design befitting an enterprise platform. Well executed on both fronts.

Prompt #6

A tube of toothpaste with the words “CYMBAL” written on it, on a bathroom counter, advertisement.

A tube of toothpaste with the words “CYMBAL” written on it, on a bathroom counter, advertisement.

As with the previous text example, Dall-E 3 surprisingly struggles with inputting the correct “CYMBAL” branding, while Imagen 2 performs accurately. This suggests inconsistency in Dall-E 3’s text generation capabilities.

Prompt #7

A mosaic-inspired portrait of a person, their features formed by a collection of small, colorful tiles.

A mosaic-inspired portrait of a person, their features formed by a collection of small, colorful tiles.

Kudos to both AI tools here — the mosaic portraits came out vibrant and compelling. I’d give a slight personal preference to Imagen 2’s composition, but both interpretations feel creative.

Prompt #8

Isometric 3D rendering of a car driving in the countryside surrounded by trees, bright colors, puffy clouds overhead.

Isometric 3D rendering of a car driving in the countryside surrounded by trees, bright colors, puffy clouds overhead.

Imagen 2 nailed the “one car” part of the prompt. Dall-E 3, on the other hand, decided to throw a carpool party, cramming four vehicles onto the scene. Also, I like Imagen 2’s vibrant palette and retro charm.

Prompt #9

A jellyfish on a dark blue background

A jellyfish on a dark blue background

A matter of stylistic taste in this case — Imagen 2 achieved photorealism with its jellyfish, while Dall-E 3 delivered a more artistic, illustrated rendition. Which one do you like better?

Prompt #10

An image of: Consider the subtleness of the sea; how its most dreaded creatures glide under water, unapparent for the most part, and treacherously hidden beneath the loveliest tints of azure

Consider the subtleness of the sea; how its most dreaded creatures glide under water, unapparent for the most part, and treacherously hidden beneath the loveliest tints of azure

The prompt is an excerpt from Moby-Dick by Herman Melville. Hence, Imagen 2 generated an abstract painting of a whale. While Dall-E 3 simply generated a random underwater scene.

Final Thoughts

Looking at the image results of Imagen 2 and Dall-E 3, I could say that the former generates more realistic and consistent images. Of course, it’s still too early to draw a conclusion since these are cherry-picked images from Imagen 2. Once the playground or API becomes available, I will do a deeper dive and write another comparison test article for you guys.

I hope this comparison gives you an idea of the differences between these AI image generators. I will also be doing a comparison with Imagen 2 and Midjourney V5, so be sure to follow and subscribe to get notified when its published.