Comparing 8 Leading AI Image Generators: Midjourney vs Stable Diffusion & More
Table of Contents
- Overview of the AI Image Generator Comparison
- Comparing Architecture Images
- Interior Design Image Comparison
- Product Design Image Results
- Character Portrait Results
- Creative Abstract Image Outcomes
- Conclusion and Favorite AI Image Generator
Overview of the AI Image Generator Comparison
I recently conducted an experiment comparing 8 different AI image generation platforms. Over the course of two days, I created 795 images across 5 different categories using the same prompts for each platform. My goal was to create a baseline to compare the outputs and see what kind of images I could generate in a short period of time.
In this blog post, I will share the results of my comparisons for each category, along with my thoughts on the strengths and weaknesses of each platform. I've included side-by-side examples from the different generators to highlight the key differences.
Image Prompts Used
Here are all of the prompts I used for generating images:
- Architectural facade
- Interior design with classical style
- Complex lighting product design
- Character portrait
- Creative abstract with neon lights
Image Generation Platforms Tested
I tested out the following 8 AI image generation platforms:
- Midjourney
- Stable Diffusion (local version)
- Blue Willow
- Lensa AI
- Dall-E 2
- Bing Image Creator
- Firefly (Adobe)
- DreamStudio
Comparing Architecture Images
For the architectural facade prompt, Midjourney and Bing Image Creator produced the most visually appealing and relevant results. The images have nicely balanced compositions with good detail in the architectural elements.
DreamStudio and Dall-E 2 were less consistent, with some of the images looking overly messy or distorted. While Stable Diffusion and Lensa AI were able to generate architectural images, they lacked some refinement compared to the top options.
Interior Design Image Comparison
The interior design prompt yielded fairly similar vintage, classical style results across most of the platforms. Blue Willow, Lensa AI, and Midjourney produced attractive, high quality interiors with nice lighting and material details.
The main differences came down to elements like the inclusion of a flying basket in Dall-E 2's image, versus more standard furniture arrangements in the other examples. Overall, this category demonstrated the AI generators can successfully produce on-theme interior design images.
Product Design Image Results
For this product lighting prompt, I was focused more on creative inspiration over photorealism. Midjourney and Dall-E 2 generated visually interesting abstract compositions that could serve as inspiration for lighting fixtures or installations.
Bing Image Creator took a more surreal approach with its eyeball lights example. Lensa AI, Stable Diffusion, and Blue Willow produced cleaner product renderings, but lacked some of the abstract creativity of the top options.
Character Portrait Results
The character portrait category yielded stunning, creative results from Midjourney and Blue Willow. The images have bold colors and dramatic lighting that bring life to the imaginary characters.
Firefly's image is nicely styled as well, though leaning more towards a cartoon style. On the other hand, Stable Diffusion's output came across as eerie and a bit unsettling.
Creative Abstract Image Outcomes
For the final creative abstract prompt, Lensa AI and Blue Willow generated the most eye-catching neon lit scenes. The images use vibrant colors, lighting effects, and abstract shapes to create stunning artistic compositions.
Stable Diffusion and Bing took a more photorealistic approach in their images. While detailed, they lacked some of the abstract creativity that the top options achieved.
Conclusion and Favorite AI Image Generator
After comparing 795 generated images across 8 platforms and 5 categories, a few key conclusions emerged:
-
Midjourney delivered the most consistent quality and creativity across categories
-
Lensa AI's custom models produce stunning abstract imagery
-
Blue Willow is nearly as good as Midjourney given its free tier
-
Bing Image Creator lags behind Midjourney but shows promise as an early Dall-E 2 demo
Taking into account factors like image quality, creativity, and value - Midjourney emerges as my top choice for an AI image generation platform. But platforms like Lensa AI and Blue Willow also have compelling strengths to offer.
FAQ
Q: Which AI image generator platforms were compared?
A: Eight leading platforms were tested: Midjourney, Blue Willow, Leonard AI, DALL-E 2, Bing Image Creator, Adobe Firefly, Dream Studio, and Stable Diffusion running locally.
Q: How were the images generated for comparison?
A: The same 5 image prompts were used to generate up to 5 images on each platform, creating 795 total images.
Q: What image categories were tested?
A: Architectural facade, interior design, product design, character portrait, and creative abstract images.
Q: Which platform performed the best overall?
A: Based on the results, Midjourney and Bing Image Creator produced the most consistent and relevant images across categories.
Q: Where can I see all the generated images?
A: Check the link in the video description for the full mood board with all 795 AI-generated images marked and labeled by platform.
Q: What features were compared?
A: Image quality, composition, creativity, style accuracy, and relevance to prompts were evaluated.
Q: Can I recreate this image generator comparison?
A: Yes, the exact prompts used are provided so you can test these platforms and compare the results yourself.
Q: Were the images edited at all?
A: No, the images shared were not edited or upscaled in any way, showing the direct generator outputs.
Q: What should I know before trying these platforms?
A: Be aware of pricing, free credits/generations offered, subscription tiers, and AI model differences on each platform.
Q: Where can I sign up for these AI image generators?
A: Check the link in the description for direct signup links and details on accessing each platform.