Embedded Files
This page displays the image results from our first test. October 2024 - Janurary 2025
Testing Process 🧪:
Testing Process 🧪:
In this experiment, we had to do some preparation before diving straight into asking the AI to generate images.
First and foremost, we did research into the best possible way to curate a proper prompt for the AI generators to interpret in an accurate way the AI will understand. We looked into multiple articles, and even asked Chat GPT for advice on creating a good prompt for testing AI image-generating websites. After getting insight from the articles and Chat GPT, we began to develop the prompts for the test, and we decided to come up with three separate prompts. The three separate prompts had different levels of complexity and needs for the image, to best showcase the generators capacity.
After making the prompts, we then made a google doc to keep track of what generators we planned to test, look into, verify, and to keep track of the exact date we evaluated each of the generators.
We then proceeded to create a new Gmail account in order to not hinder the process of the testing on image generation since some image generators take into account from the user's previous history of usage which can influence the results of the image. While we inputted the the prompt one by one in the text box, at the same time, we used a google form filled with detailed questions which helped us compare and contrast not only the ability of how well the generator produces imaged but it's additional functionalities, features and price (accessibility) which are other important aspects when using a image generating website.
🚨 *DISCLAIMERS* 🚨
🚨 *DISCLAIMERS* 🚨
*DISCLAIMER* Results of rankings may be skewed due to the different number of images with each prompt.
ex: Midjourney returns four images while DALL-E provides only a singular image each prompt, I pick the best image out of the four Midjourney outputs/generates.
Some generators had a text limit or had a specific feature where you must pick a specific style, rather than typing it into the prompt so results maybe very skewed, especially Stable Diffusion.
Some generators use the same models; outputs may therefore end up with similar compositions.
The rankings on this page are determined by the accumulation of points from each respective prompt using the point system found in the "Prompt x Ranking" pages (button link to pages down below).
The format of the results are ordered by the time it was tested from first to last, not by rank.
Using the buttons below, go to the specific prompt pages in order to compare how each AI software produces material from the same prompt.
PROMPT 1:
PROMPT 1:
PROMPT 2:
PROMPT 2:
PROMPT 3:
PROMPT 3:
An adult girl with brown hair wearing a hiking uniform leaning against a tree at the top of a hill in the middle of the forest overlooking an abandoned building that is a few floors tall in the distance. The scene takes place during the season of fall in the middle of a sunny day. The scene takes place during the fall season, with vibrant orange and red leaves scattered around. Sunlight filters through the trees, casting dappled shadows on the ground. The style should be realistic but with a slightly cartoony touch, similar to Disney's art style.
Two serious-looking adult men in tailored suits playing a card game, sitting at a small round table, opposite of each other but facing each other in the center of a dimly lit room. The room has minimal furnishings, with a vintage hanging lamp casting shadows across the scene. The image should be a simple, less detailed black and white illustration in the style of mid-20th-century newspaper cartoons, with bold lines and exaggerated facial expressions.
An underwater aquarium filled with a variety of marine life, including sleek sharks, colorful tropical fish, and playful dolphins. The scene is bathed in a soft blue light, creating a dreamy and calming atmosphere. On the left side of the image, a child with a relaxed posture, stands close to the glass, watching the aquatic animals with a gentle smile. The right side of the image shows the marine creatures swimming gracefully in their enclosure, with light reflecting off the water creating a shimmering effect making the right side of the image more vibrant compared to the left side which will contrast by being less vibrant. The overall color palette should be cool and soothing, dominated by shades of blue and teal.
Release date: January 2021
Date of Test: October 2024
DALL-E Is a GPT (GPT stands for a Generative Pre-trained Transformer, a type of language model that uses deep learning to create human-like text) developed by OpenAI which is Chat GPT's dedicated image generator so it can only be accessed and used within the Chat GPT website or app since it's only implemented there.
Rank 3
DALL-E SPECS
DALL-E SPECS
Is there a free version?
Yes but limited to 3 entries per day.
Type of input allowed:
Text
Number of images returned per prompt:
1
Does this AI support API?
Yes
File type of image download: which type(s) does the AI support:
.webp file
Supports Infill: (regenerating only a selected part of an image)
Yes
Does the AI support Out fill (Extending the size of the regenerated image?
Yes
Does the AI support upscale (making a generated image larger and more detailed)?
Yes
Style: Does the AI support different styles? (realistic, abstract, anime, etc.)?
Yes
Aspect Ratio: Can you control the aspect ratio of the image?
Not sure
Image Quality:
Excellent quality, professional-grade
Creativity and Flexibility:
Exceptional creativity and customizability
AI Capabilities and Control:
Effective iteration and refinement
Ethical and Legal Considerations: How transparent is the AI about IP and data usage?
Adequate IP rights and disclosures
Additional Comments: What is your overall opinion of this AI platform?
After viewing the results of the three prompts, I believe DALL-E had done an exceptional job generating each image, since they were not too far from what I've expected. The pictures felt stunning and had more of a human touch with great composition to it than I expected. I would rate it a 9/10.
Release Date: July 2022.
One of the most popular AI image generating tool, alongside with DALL-E and Stable diffusion. The founder is a man named David Holz. This website is the most polished in terms of image editing functions and user interface.
Release Date: July 2022.
Rank 2
Midjourney SPECS
Midjourney SPECS
Is there a free version?
No, base or lowest price is $8 or $10 monthly.
Type of input allowed:
Text, Image
Number of images returned per prompt:
4
Does this AI support API?
No
File type of image download: which type(s) does the AI support:
.png file
Supports Infill: (regenerating only a selected part of an image)
Yes
Does the AI support Out fill (Extending the size of the regenerated image?
Yes
Does the AI support upscale (making a generated image larger and more detailed)?
Yes
Style: Does the AI support different styles? (realistic, abstract, anime, etc.)?
Yes
Aspect Ratio: Can you control the aspect ratio of the image?
Yes: one can explicitly request a specific aspect ratio
Image Quality:
Excellent quality, professional-grade
Creativity and Flexibility:
Versatile in adapting to different styles
AI Capabilities and Control:
Exceptional creative autonomy
Ethical and Legal Considerations: How transparent is the AI about IP and data usage?
Adequate IP rights and disclosures
Additional Comments: What is your overall opinion of this AI platform?
The interface is really satisfying and polished, the generation time takes a lot longer than most generation sites. Surprisingly it doesn't follow the prompts when I tell it to be less realistic, but follows more prompt instructions compared to the others.
Release Date: June 2023
Adobe Firefly, is an AI tool integrated into Photoshop and Illustrator to streamline project edits, saving time by automating fixes. It also offers a standalone website with advanced image generation features.
Rank 6
Adobe Firefly SPECS
Adobe Firefly SPECS
Is there a free version?
No Adobe subscription is needed but if you are a college student, technically it's free.
Type of input allowed:
Text
Number of images returned per prompt:
3
Does this AI support API?
Yes
File type of image download: which type(s) does the AI support:
.png
Supports Infill: (regenerating only a selected part of an image)
Yes
Does the AI support Out fill (Extending the size of the regenerated image?
No
Does the AI support upscale (making a generated image larger and more detailed)?
Yes
Style: Does the AI support different styles? (realistic, abstract, anime, etc.)?
Yes
Aspect Ratio: Can you control the aspect ratio of the image?
Yes: one can explicitly request a specific aspect ratio
Image Quality:
High clarity, good resolution
Creativity and Flexibility:
Versatile in adapting to different styles
AI Capabilities and Control:
Moderate response to complex prompts
Ethical and Legal Considerations: How transparent is the AI about IP and data usage?
Exemplary ethical standards
Additional Comments: What is your overall opinion of this AI platform?
It does not generate faces properly. Some of the faces would be nightmare fuel but it follows the prompts pretty decently.
Release Date: 2016
Developed by Deep AI Inc., it is one of the first company to offer online AI text to image generator.
Rank 4
DeepAI SPECS
DeepAI SPECS
Is there a free version?
Yes but limited.
Type of input allowed:
Text
Number of images returned per prompt:
1
Does this AI support API?
Yes
File type of image download: which type(s) does the AI support:
.jfif
Supports Infill: (regenerating only a selected part of an image)
No
Does the AI support Out fill (Extending the size of the regenerated image?
No
Does the AI support upscale (making a generated image larger and more detailed)?
No
Style: Does the AI support different styles? (realistic, abstract, anime, etc.)?
Yes
Aspect Ratio: Can you control the aspect ratio of the image?
Yes: one can explicitly request a specific aspect ratio
Image Quality:
Moderate clarity but needs refinement
Creativity and Flexibility:
Versatile in adapting to different styles
AI Capabilities and Control:
Moderate response to complex prompts
Ethical and Legal Considerations: How transparent is the AI about IP and data usage?
N/A
Additional Comments: What is your overall opinion of this AI platform?
The quality was worse. Although it didn't follow the instructions as properly as I wanted to since it didn't make the girl lean on the tree as I told it to in the prompt. It did follow the instructions in terms of where I wanted the subjects to be placed in the 3rd prompt where I wanted the child to be placed on the left side.
Rank 10
Art Breeder SPECS
Art Breeder SPECS
Is there a free version?
Yes
Type of input allowed:
Text, Image
Number of images returned per prompt:
More than 4
Does this AI support API?
Yes
File type of image download: which type(s) does the AI support:
.png
Supports Infill: (regenerating only a selected part of an image)
No
Does the AI support Out fill (Extending the size of the regenerated image?
No
Does the AI support upscale (making a generated image larger and more detailed)?
No
Style: Does the AI support different styles? (realistic, abstract, anime, etc.)?
Yes
Aspect Ratio: Can you control the aspect ratio of the image?
Yes: one can explicitly request a specific aspect ratio
Image Quality:
High clarity, good resolution
Creativity and Flexibility:
Limited output diversity: How diverse are the outputs?
AI Capabilities and Control:
Basic prompt understanding
Ethical and Legal Considerations: How transparent is the AI about IP and data usage?
N/A
Additional Comments: What is your overall opinion of this AI platform?
Image generation took a long time when I turned on the "realistic mode" but I switched it over to the regular one. The images didn't follow the prompts that well, and the style was very limited. Each image generated looked the same as the other. It even went off what I was trying to convey in the prompt. I think Art Breeder is mostly good for portraits of people of different ethnicities and varying their looks rather than specific image generation.
NightCafe SPECS
NightCafe SPECS
Is there a free version?
Yes but limited.
Type of input allowed:
Text, Image
Number of images returned per prompt:
1
Does this AI support API?
Unknown
File type of image download: which type(s) does the AI support:
.jpg
Supports Infill: (regenerating only a selected part of an image)
No
Does the AI support Out fill (Extending the size of the regenerated image?
No
Does the AI support upscale (making a generated image larger and more detailed)?
No
Style: Does the AI support different styles? (realistic, abstract, anime, etc.)?
No
Aspect Ratio: Can you control the aspect ratio of the image?
Yes: one can explicitly request a specific aspect ratio
Image Quality:
High clarity, good resolution
Creativity and Flexibility:
Versatile in adapting to different styles
AI Capabilities and Control:
Effective iteration and refinement
Ethical and Legal Considerations: How transparent is the AI about IP and data usage?
Adequate IP rights and disclosures
Additional Comments: What is your overall opinion of this AI platform?
There were a lot of annoying popups, and had annoying interface which made it off-putting to use. It didn't follow the prompt too well nor did it stray off too badly. I'd give it a 7/10
Release date: February 2023
Founders are Clément Delangue, Julien Chaumond, and Thomas Wolf. Hugging Face's image generation tools thrive in an open-source ecosystem, offering access to a variety of pre-trained models, including cutting-edge diffusion-based models like Stable Diffusion, which produce high-quality and photorealistic images from text prompts. Users can fine-tune these pre-trained models or train their own, enabling personalized outputs and creative flexibility.
Rank 9
Hugging Face SPECS
Hugging Face SPECS
Additional Information:
Additionally, Hugging Face provides comprehensive educational resources, including documentation, tutorials, and examples, making it easy for users of all levels to experiment and innovate with image generation technologies.
Is there a free version?
Yes
Type of input allowed:
Text
Number of images returned per prompt:
1
Does this AI support API?
Yes
File type of image download: which type(s) does the AI support:
.jpg
Supports Infill: (regenerating only a selected part of an image)
No
Does the AI support Out fill (Extending the size of the regenerated image?
No
Does the AI support upscale (making a generated image larger and more detailed)?
No
Style: Does the AI support different styles? (realistic, abstract, anime, etc.)?
No
Aspect Ratio: Can you control the aspect ratio of the image?
No
Image Quality:
Low resolution, unclear details
Creativity and Flexibility:
Limited output diversity
AI Capabilities and Control:
Basic prompt understanding
Ethical and Legal Considerations: How transparent is the AI about IP and data usage?
Adequate IP rights and disclosures
Additional Comments: What is your overall opinion of this AI platform?
Didn't follow prompts style, not user friendly, probably only good for realistic styles.
Release date: March 2023
Microsoft CoPilot is integrated into the Microsoft ecosystem, like Word, PowerPoint, Excel, and other Office 365 applications. By leveraging advanced AI models, such as OpenAI’s DALL-E, it enables users to generate custom images directly within these applications. It also has a stand alone website.
Rank 8
Prompt 1 unable to be generated in Copilot due to unknown restrictions
Microsoft Copilot SPECS
Microsoft Copilot SPECS
Is there a free version?
Yes
Type of input allowed:
Text
Number of images returned per prompt:
2
Does this AI support API?
Yes
File type of image download: which type(s) does the AI support:
.jpg
Supports Infill: (regenerating only a selected part of an image)
No
Does the AI support Out fill (Extending the size of the regenerated image?
No
Does the AI support upscale (making a generated image larger and more detailed)?
No
Style: Does the AI support different styles? (realistic, abstract, anime, etc.)?
Yes
Aspect Ratio: Can you control the aspect ratio of the image?
Not sure
Image Quality:
Excellent quality, professional-grade
Creativity and Flexibility:
Versatile in adapting to different styles
AI Capabilities and Control:
Effective iteration and refinement
Ethical and Legal Considerations: How transparent is the AI about IP and data usage?
Adequate IP rights and disclosures
Additional Comments: What is your overall opinion of this AI platform?
It didn't let me use my first prompt due to it crossing the guidelines, didn't tell me what word in my prompt crossed the guidelines even when I tried to adjust it as much as possible, it didn't go through. It's still pretty good nonetheless 9/10.
Release date: August 2, 2024
Developed by developed by Chengdu Yiwo Tech Development Co., Ltd.
Rank 7
PicLumen SPECS
PicLumen SPECS
Is there a free version?
Yes
Type of input allowed:
Text
Number of images returned per prompt:
4
Does this AI support API?
Yes
File type of image download: which type(s) does the AI support:
.jpg, .png
Supports Infill: (regenerating only a selected part of an image)
No
Does the AI support Out fill (Extending the size of the regenerated image?
No
Does the AI support upscale (making a generated image larger and more detailed)?
No
Style: Does the AI support different styles? (realistic, abstract, anime, etc.)?
Yes
Aspect Ratio: Can you control the aspect ratio of the image?
Yes: one can explicitly request a specific aspect ratio
Image Quality:
High clarity, good resolution
Creativity and Flexibility:
Versatile in adapting to different styles
AI Capabilities and Control:
Effective iteration and refinement
Ethical and Legal Considerations: How transparent is the AI about IP and data usage?
Adequate IP rights and disclosures
Additional Comments: What is your overall opinion of this AI platform?
It's as good as Deep ai which is a good option for free ai image generation. It's catered towards anime style mostly.
Release date: 2022
Google Labs is a platform for users to interact with new, experimental AI tools developed by Google, particularly models like LaMDA (Language Model for Dialogue Applications). The Test Kitchen gives users a chance to experiment with various AI technologies and offer direct feedback to help improve the models.
Rank 1
Google Labs SPECS
Google Labs SPECS
Additional Information:
This initiative started as a way to fine-tune LaMDA 2, which was introduced at Google I/O 2022, and then expanded to other models such as Imagen, a text-to-image AI. AI Test Kitchen serves as a way for Google to gather insights about user experiences with its cutting-edge technologies before making them more widely available.
Is there a free version?
Yes, but limited
Type of input allowed:
Text
Number of images returned per prompt:
4
Does this AI support API?
Unknown
File type of image download: which type(s) does the AI support:
.jpg
Supports Infill: (regenerating only a selected part of an image)
Yes
Does the AI support Out fill (Extending the size of the regenerated image?
No
Does the AI support upscale (making a generated image larger and more detailed)?
No
Style: Does the AI support different styles? (realistic, abstract, anime, etc.)?
Yes
Aspect Ratio: Can you control the aspect ratio of the image?
Yes, but limited or must included in the prompt
Image Quality:
High clarity, good resolution
Creativity and Flexibility:
Versatile in adapting to different styles
AI Capabilities and Control:
Effective iteration and refinement
Ethical and Legal Considerations: How transparent is the AI about IP and data usage?
Exemplary ethical standards
Additional Comments: What is your overall opinion of this AI platform?
It's one of the top ai image generator
Release Date: November 2023
Grok is an AI chatbot developed by Elon Musk's company X. It's designed to be truthful, curious, and useful, and to answer questions in a witty and rebellious way. Grok is available on X (formerly Twitter), its own website, and an iOS app.
Rank 5
Grok 2 SPECS
Grok 2 SPECS
Is there a free version?
Yes, but limited
Type of input allowed:
Text, Image
Number of images returned per prompt:
4
Integration: Is this AI standalone or integrated into another application? (for example, within Adobe Photoshop or Google Slides) If so, what application?
Grok is integrated into X (formally known as twitter) so you need to create an account with X in order to access it.
Does this AI support API?
Yes
File type of image download: which type(s) does the AI support:
.jpg
Supports Infill: (regenerating only a selected part of an image)
No
Does the AI support Out fill (Extending the size of the regenerated image?
No
Does the AI support upscale (making a generated image larger and more detailed)?
No
Style: Does the AI support different styles? (realistic, abstract, anime, etc.)?
Yes
Aspect Ratio: Can you control the aspect ratio of the image?
No
Image Quality:
Moderate clarity but needs refinement
Creativity and Flexibility:
Some style flexibility
AI Capabilities and Control:
Moderate response to complex prompts
Ethical and Legal Considerations: How transparent is the AI about IP and data usage?
Partially transparent about data usage
Additional Comments: What is your overall opinion of this AI platform?
Not the greatest image generator compared to the top AI.
Release Date: August 22, 2022
Stable Diffusion was developed by Stability AI, an AI-focused company founded by Emad Mostaque in 2020. Stability AI aims to democratize AI technologies and make them accessible for both individuals and businesses worldwide.
Rank 11
Stable Diffusion SPECS
Stable Diffusion SPECS
Additional Information:
Stable Diffusion is an open-source AI model that has gained widespread adoption due to its accessibility and flexibility. It uses a latent diffusion process to efficiently generate high-resolution and detailed images, even on consumer-grade GPUs. Its open-source nature has fostered a vibrant community, with contributors continuously developing extensions and tools to enhance its functionality. Additionally, users can fine-tune the model to create art in specific styles or themes, making it highly versatile for various creative needs.
Is there a free version?
Yes, but limited
Type of input allowed:
Text, Image
Number of images returned per prompt:
4
Does this AI support API?
Yes
File type of image download: which type(s) does the AI support:
.jpg
Supports Infill: (regenerating only a selected part of an image)
No
Does the AI support Out fill (Extending the size of the regenerated image?
No
Does the AI support upscale (making a generated image larger and more detailed)?
Yes
Style: Does the AI support different styles? (realistic, abstract, anime, etc.)?
Yes
Aspect Ratio: Can you control the aspect ratio of the image?
Yes: one can explicitly request a specific aspect ratio
Image Quality:
Excellent quality, professional-grade
Creativity and Flexibility:
Some style flexibility
AI Capabilities and Control:
Moderate response to complex prompts
Ethical and Legal Considerations: How transparent is the AI about IP and data usage?
Adequate IP rights and disclosures
Additional Comments: What is your overall opinion of this AI platform?
If you wanted to have a cartoon styled image you must specifically pick the style choice options feature, it won't pick it up in the prompt, but for the sake of this experiment I didn't use the feature. (I think to get the most effective results from certain generators you must pick specific style options that match your request or else it won't give you the style you want) There is another feature which is a sliding bar that adjusts to how much you want it to follow the prompt, I left it in it's default setting which was 7 which was pretty low on the bar but the results maybe better if I were to adjust it to a higher number. Lastly, there is a forced watermark in the bottom right side, and in order to remove it you need to pay, this is the first time I've seen a generator use their specific watermark.
Testing, Research and Website Design by Priya Begum as of September 2024 - January 28, 2025
Page updated
Google Sites
Report abuse