Making mazes with AI: Google Gemini - Has it improved over time ?

Welcome back to my series of posts where I test AI image generators and see how they handle making maze art. I will be asking 10 prompts and seeing what gets generated. My goal is to evaluate different AI image sites against each other to see how they perform. Here is the series so far:

An exploration into Al Image Maze Generation

Making mazes with AI: DallE2 (Updated to compare 2023 to 2026)

Making mazes with AI: Stable Diffusion (Updated to compare 2023 and 2024 to 2026)

Making mazes with AI: Craiyon (Updated to compare 2023 to 2026)

Making mazes with AI: Dream by Wombo (Updated to compare 2023 to 2026)

Making mazes with AI: Nightcafe (Updated to compare 2023 to 2026)

Making mazes with AI: StarryAI (Updated to compare 2023 to 2026)

Making mazes with AI: DeepAI (Updated to compare 2023 to 2026)

Making mazes with AI: Canva (Updated to compare 2023 and 2024 to 2026)

Making mazes with AI: AI Image Generator (Updated to compare 2023 to 2026)

Making mazes with AI: ToolBaz

Making mazes with AI: Photosonic

Making mazes with AI: Bing AI Image Generator (Updated to compare 2023 and 2024 to 2026)

 
Gemini logo.

Making Maze Art with google gemini

2024: You can access the Gemini website here. You must sign into your google account to use Gemini. There is not limit to the number of images you can generate. Each prompt returns 2 images with the option to generate more. For this exercise I chose the one closest to what I had asked for, or the most interesting.

My original testing took place in March 2023. Gemini (formerly Bard) just started this feature in Feb 2024 so it late to the party.

2026: You still need to sign in to use the site. Just type and the prompt will use Non Banana to create an image for you.

 

Prompt 1 - Make a medium difficulty maze of the Eiffel Tower in black and white with arrows at the start and finish

2024: That is the Eiffel Tower. There is a maze looking section and an arrow. Typically I do not get arrows, so this is an improvement.

2026: When this first generated and I saw the arrows and the maze inside the Eiffel Tower I thought we had out first AI maze art generation…but not quite.

 

Prompt 2 - Draw a medium difficulty large maze of the Empire State Building with the start and goal embedded in the structure

2024: Similar to the 1st maze but in a new style. Got the famous building correct. Looks like a maze but is not a maze.

2026: So close again. The start and goal make no sense no I see the building and the maze.

 

Prompt 3 - Draw a difficult maze of the White House pixel art style

2024: Again, we have the right building. Wrong style. No Maze. But good job if I had written a different prompt.

2026: Similar to the previous new generations, the start and goal make no sense, but the building is correct and I see an internal maze !

 

Prompt 4 - Draw a difficult maze that looks like a drawing of a famous building in sketch style

2024: The top is a sketch of the very famous ______ building. The bottom is a hand drawn maze that actually is a maze, i mean if it had a start and goal !

2026: Great sketch of the Roman Coliseum in Rome with no maze, but a start and a goal.

 

Prompt 5 - Draw a maze in the style of doyoumaze.com of a skyscraper in NYC

2024: Cool NYC image. I think I like the marble sky. No Maze, not my style, but at least it is cool !

2026: Reminds me of the ESB above.

 

Prompt 6 - Draw a maze in the style of Sean C Jackson of a scene from a large outdoor market

2024: This has potential. Change the perspective and switch from watercolor to more comic and you’ve got it.

2026: It is a fun comic style black and white market using circus tents. Has potential. Crowded.

 

Prompt 7 - Make a maze of a slice of an orange in color

2024: Sure.

2026: The start and finish make sense and the random arrow on the right side could be removed.

 

Prompt 8 - Make a maze integrated on top of a photograph of a king sitting on his throne looking cantankerous beside his beautiful queen

2024: Weird. I didn’t ask for a specific king and queen. Any would do. If I ask for a plumber or bus driver is it blocked ? And, after I asked this…well you all saw the many stories !

2026: An overlay of an (almost) maze over a photo…I guess Google figured out that this prompt was fine.

 

Prompt 9 - Make a solvable maze that is very large and very difficult to solve because it is so complex

2024: Looks like a hedge maze from above. Nice style.

2026: A maze algorithm away from being great.

 

Prompt 10 - Make a 3d render of a red and blue glossy cube maze

2024: No.

2026: The old double start.

 

Summary and Conclusion:

2024: How did Google Gemini do ? I think it was average. It has less structure than other text to image generators. You can ask for whatever you want (except people I guess). And there are no limits. But it also has no built in styles or adjustment buttons that are helpful with other AI image sites.

2026: Although no generator has incorporated maze algorithms into their process I think Gemini fakes it best. Solid output across the board and one of the reasons I use Gemini most often when I need a background or to make a comic strip.

Making mazes with AI: Bing AI Image Generator - Has it improved over time ?

Welcome to the 13th in a series of posts where I will test AI image generators and see how they handle making maze art. I will be asking 10 prompts and seeing what gets generated. My goal is to evaluate different AI image sites against each other to see how they perform. In my first post I discussed the project, and today we start with the first site evaluation: Bing AI Image Generator which is powered by DallE. Here is the series so far:

An exploration into Al Image Maze Generation

Making mazes with AI: DallE2 (Updated to compare 2023 to 2026)

Making mazes with AI: Stable Diffusion (Updated to compare 2023 and 2024 to 2026)

Making mazes with AI: Craiyon (Updated to compare 2023 to 2026)

Making mazes with AI: Dream by Wombo (Updated to compare 2023 to 2026)

Making mazes with AI: Nightcafe (Updated to compare 2023 to 2026)

Making mazes with AI: StarryAI (Updated to compare 2023 to 2026)

Making mazes with AI: DeepAI (Updated to compare 2023 to 2026)

Making mazes with AI: Canva (Updated to compare 2023 and 2024 to 2026)

Making mazes with AI: AI Image Generator (Updated to compare 2023 to 2026)

Making mazes with AI: ToolBaz

Making mazes with AI: Photosonic

Update from 2026:

When I did this series in Spring of 2023 I found that none of the text to image generators were able to generate actual mazes, including this one. They were able to make images that looked maze-like, but none were solvable puzzles. The 10 prompts below are a record of what was possible then, but each generation of AI pushes the limit further on what is possible. I remade each maze in 2024 and now in 2026 for comparison. All of these were made using text to image generation and not coding.

Bing AI logo

Making Maze Art with BING AI Image generator

2023: You can access the Bing image generator website here. You must sign in to your Microsoft account. As of this post there is no limit to the number of prompts you can ask for, and it is free to use. Each prompt, or generation from will generate 4 image options. For this exercise I chose the one closest to what I had asked for, or the most interesting.

2026: Everything seems similar to me in 2026. You still need to sign-in to your account. 4 images generated.

Note: Because each comparison includes 3 images, if you want to see a larger version just clock on it.

 

Prompt 1 - Make a medium difficulty maze of the Eiffel Tower in black and white with arrows at the start and finish

2023: I see the Eiffel Tower. This is a really cool interpretation for a maze. The other options were more traditional, but this one is so interesting ! I even see an arrow of some sort at the top above the tower.

2024: The 2024 version includes trees and a boulevard rather than an abstract maze. It also includes a more detailed version of the tower. Has a maze in the sky and on the ground. 2024 is an improvement.

2026: I see a maze drawn on the Eiffel Tower. Very cool.

 

Prompt 2 - Draw a medium difficulty large maze of the Empire State Building with the start and goal embedded in the structure

2023: Wow. Such an interesting take on a maze of the ESB. All 4 options were good.

2024: In the 2024 version I like the maze being built into the buildings of NYC around the Empire State Building which has a more accurate shape than the 2023 version. Nice improvement !

2026: What is that ? Very cool surrounding maze concept but a huge miss on the ESB,

 

Prompt 3 - Draw a difficult maze of the White House pixel art style

bing ai white house 2026: It is pixel art style. It has a white house in the center of the image but not really the White House.

2024: 2024 is excellent although I do not understand the god view of the artist with the desk.

2026: Looks like something I might build in a video game.

 

Prompt 4 - Draw a difficult maze that looks like a drawing of a famous building in sketch style

2023: This IS a sketch of a building maze, just not a famous building. Still I like how it came out !

2024: I do not know what building 2024 is….but it is very well done ! Cool maze like structure in the front of the building.

2026: I am going to guess that this one is Greek. Maze-like design on the roof.

 

Prompt 5 - Draw a maze in the style of doyoumaze.com of a skyscraper in NYC

2023: None of the 4 was in my style (in fact each had a different style to it), but this is the closest. Could easily be for the Empire State Building prompt above.

2024: The Chrysler building in a circular maze frame. Nice artwork.

2026: This looks like a fortress from a city in the future. Cool maze-like drawings on the side.

 

Prompt 6 - Draw a maze in the style of Sean C Jackson of a scene from a large outdoor market

2023: This is really good. The style is a bit off but it captured what was asked for better than anything.

2024: I see some of the SCJ style in 2024.

2026: I think I prefer 2026. So much detail in the market and it is at the correct angle.

 

Prompt 7 - Make a maze of a slice of an orange in color

2023: I wasn’t sure what I was looking for. I think this is it !

2024: 2024 is interesting with some of the oranges wedges being stuffed with other fruits and veggies.

2026: Looks like an artist made it. Very interesting and could be further developed into a real maze.

 

Prompt 8 - Make a maze integrated on top of a photograph of a king sitting on his throne looking cantankerous beside his beautiful queen

2023: Weird. Be specific in what you are asking for !

2024: You can see the improvement so much in this prompt with how the king and queen were rendered.

2026: And even more improvement into 2026. We go from a glare from the queen to a king looking away to a queen away over the year. Couples counseling is needed.

 

Prompt 9 - Make a solvable maze that is very large and very difficult to solve because it is so complex

2023: I like the idea of this guy lost in a maze where he is so tall he should be able to see his way out ? Am I wrong about this ?

2024: I like the use of small people using ladders around a maze to get around.

2026: This is large, complex and cool looking. Great job if it was an actual maze.

 

Prompt 10 - Make a 3d render of a red and blue glossy cube maze

2023: 4 shown because all were interesting and none was better than another. None are a maze. Glossy or Shiny ?

2024: 4 to choose from. Mazelike cubes.

2026: Interesting. It loses some of the red and blue, but I could imagine a maze inside. Similar to 2024.

 

Summary and Conclusion:

2023: How did Bing AI Image Generator do ? I think it had a strong showing. I expect it do do well against the other AI’s. It showed it can use a variety of styles and never got tripped up completely with any of the prompts. I will use this again for sure !

2026: I think it improved. More interesting generations, with more detail, and better adherence to what is asked in the prompts. Bing AI is one of the 3 text to image generators I use depending on what I need (Gemini and starryai are the other 2) and I am impressed with how Bing AI did ! I expect this to get even better in the future !

Coming next: Comparison of the 12 AI generating websites

Making mazes with AI: AI Image Generator - Has it improved over time ?

Welcome to the 10th in a series of posts where I will test AI image generators and see how they handle making maze art. I will be asking 10 prompts and seeing what gets generated. My goal is to evaluate different AI image sites against each other to see how they perform. If you need to catch-up here is the series so far:

An exploration into Al Image Maze Generation

Making mazes with AI: DallE2 (Updated to compare 2023 to 2026)

Making mazes with AI: Stable Diffusion (Updated to compare 2023 and 2024 to 2026)

Making mazes with AI: Craiyon (Updated to compare 2023 to 2026)

Making mazes with AI: Dream by Wombo (Updated to compare 2023 to 2026)

Making mazes with AI: Nightcafe (Updated to compare 2023 to 2026)

Making mazes with AI: StarryAI (Updated to compare 2023 to 2026)

Making mazes with AI: DeepAI (Updated to compare 2023 to 2026)

Making mazes with AI: Canva (Updated to compare 2023 and 2024 to 2026)

Update from 2026:

When I did this series in Spring of 2023 I found that none of the text to image generators were able to generate actual mazes, including this one. They were able to make images that looked maze-like, but none were solvable puzzles. The 10 prompts below are a record of what was possible then, but each generation of AI pushes the limit further on what is possible. I remade each maze in 2026 for comparison. All of these were made using text to image generation and not coding.

AI Image Generator logo

Making Maze Art with AI Image Generator

2023: Today I test out AI Image Generator. You can access the AI Image Generator website here. At the time of this writing it is ranked #1 on Google search for AI image generation (makes sense with that name). You can just go to the website and start using it , there is no sign-in or anything. Each prompt, or generation will generate 1 image. I found that the site would error out a lot, and during my 10 prompts I got about an equal number or errors. All prompts I use are the same for each site !

2026: Still no sign-in required. A basic looking site that is very type and go. I found the generations to go smoothly this time, with no errors like in 2023. There is a prompt options at the bottom you can select 3 different topics: Art Style (with 10 options); Subject (with 6 options); and Mood & Lighting (with 8 options). You can also upload an image for guidance. I used the default for all, with 1 image generated each time.

Prompt 1 - Make a medium difficulty maze of the Eiffel Tower in black and white with arrows at the start and finish

2023: It is almost a maze map of the city surrounding the Eiffel Tower which is a really nice idea, but it isn’t quite that.

2026: I like the white drawing on black. Not quite a maze but much improvement from 2023.

 

Prompt 2 - Draw a medium difficulty large maze of the Empire State Building with the start and goal embedded in the structure

2023: I see the base of the ESB and the core…with the top, actually 2 tops on each side of the building…

2026: I see a line drawing of the ESB. The surrounding buildings are weak and there is no maze but still this is a massive improvement.

 

Prompt 3 - Draw a difficult maze of the White House pixel art style

2023: Nope.

2026: Much better. We have a recognizable building and a maze like drawing on the lawn. The red flag is awkward. Just leave it white if you are going to color it incorrectly.

 

Prompt 4 - Draw a difficult maze that looks like a drawing of a famous building in sketch style

2023: I like the angles in the mazelike drawing that looks nothing like any famous building I can think of.

2026: I mean at least this time we have a building surrounded by a maze.

 

Prompt 5 - Draw a maze in the style of doyoumaze.com of a skyscraper in NYC

2023: Almost a maze. Not close to NYC or really a building either.

2026: This would be a really nice generation for my ESB prompt above, but it also fits here of course.

 

Prompt 6 - Draw a maze in the style of Sean C Jackson of a scene from a large outdoor market

2023: No market. It looks like a maze within a labyrinth which is a fun idea !

2026: Nice looking market with weird mazelike flower boxes. Better try but a miss none the less.

 

Prompt 7 - Make a maze of a slice of an orange in color

2023: I have never seen a slice of oranges cut into a square piece before.

2026: An orange slice with a very awkward circular center.

 

Prompt 8 - Make a maze integrated on top of a photograph of a king sitting on his throne looking cantankerous beside his beautiful queen

2023: Not even close.

2026: Hey now. A nice job of doing what was asked.

 

Prompt 9 - Make a solvable maze that is very large and very difficult to solve because it is so complex

2023: So large is this maze that it must be completely off the page. Is that a small man standing in the middle ?

2026: The maze is easy but it looks cool.

 

Prompt 10 - Make a 3d render of a red and blue glossy cube maze

2023: Not bad. Not good either.

2026: Interesting.

 

Summary and Conclusion:

2023: How did AI Image Generator do ? Once you get past all the errors it made some interesting concepts that could be used as inspiration for other mazes. I would not recommend it otherwise.

2026: This is on the low end of the generators I’ve tested but it is still much improved. It erred on the side of mazes in 2023, and images without a maze in 2026. The 2026 versions are much better and closer to what was asked for. I would still not recommend it.

Coming Next: ToolBaz

One Maze - Many designs - experiment #7

This is a continuation of of maze design experiments. Read about it in post #1, One Maze - Many designs, then The first 5 experiments and so on, or just enjoy the ride.

The original maze:

Simple Maze

I used Inkscape for all of these creations. All of these were me playing with the borders and making them in different patterns and shapes'.

  • The first maze I used red bubble walls. It is fine.

  • The second uses jagged walls that are uniform in their back and forth-ness. It would be interesting to make this more random and see how it looks.

  • The third maze features curved half lines, half moons.

  • The fourth maze is diamond walls, just because.

  • The fifth random squiggle walls that are repeating. I created a free hand shape and then repeated it as a wall. Fun. I like this one best.

I did a lot of copying and pasting to create these mazes. None of these are automated in Inkscape. I’m sure there is software that can do this for you after you load a shape.