My overall verdict on the results of the tests

This page has 0 threads | Add post

Blogs

Many people around the world enjoy and benefit from our blogs. If you're one of them - and can afford to - please make a small donation to help keep this and future blogs, exercises, skills assessment tests, tips, tutorials and videos freely available to everyone!

Some other pages relevant to these blogs include:

AI (Artificial Intelligence) training courses

You can also book hourly online consultancy for your time zone with one of our 7 expert trainers!

Recomparing the main AI image creation tools 2 years on Part five of a five-part series of blogs
In this blog I re-run AI image creation comparison tests I performed way back in July 2024 to see how the results - and the tools which produce them - have changed. Recomparing the main AI image creation tools 2 years on First test - photo-realism Second test - following instructions Third test - being creative My overall verdict on the results of the tests (this blog)

Recomparing the main AI image creation tools 2 years on

Part five of a five-part series of blogs

In this blog I re-run AI image creation comparison tests I performed way back in July 2024 to see how the results - and the tools which produce them - have changed.

Recomparing the main AI image creation tools 2 years on
First test - photo-realism
Second test - following instructions
Third test - being creative
My overall verdict on the results of the tests (this blog)

Posted by Andy Brown on 18 May 2026

In this blog

In 4th place
In 3rd place
Joint first place

I've run quite a few of these AI comparison tests now, and this is probably the easiest one I've done to judge.

In 4th place

Firmly in last place was Claude, whose answer to every image creation problem seems to involve cartoons (I should have realised this when I compared the general performance of the AI tools in this blog, and Claude drew a cartoon version of a travel bingo card).

The best two of Claude's pictures.

If I were Anthropic I would withdraw this part of Claude from the market until it gives at least adequate results.

In 3rd place

I thought Copilot's images were consistently weak:

The best two of Copilot's images.

In addition, Copilot was by a long way the slowest tool, and came to an almost complete standstill trying to generate a hangman's gibbet.

I normally don't take too much account of image generation time, as I tend to set an AI tool going and then come back to see the results after doing other tasks, but Copilot really did take an unacceptably long time to produce anything.

Joint first place

I thought there wasn't much to choose between ChatGPT and Gemini. ChatGPT's images were possibly slightly better:

The best two of ChatGPT's images.

However, Gemini managed to produce its images typically in half the time:

The best two of Gemini's images.

I will definitely choose between these two tools for my image generation for the coming months.

If I were the other AI tool providers I would be very worried about how much progess Gemini has made in a very short time. Two years ago neither Claude nor Gemini had an image creation facility, but they have made unequal progress. Google use their own chips and data centres for AI, which is a policy clearly paying dividends.

Parts of this blog
Recomparing the main AI image creation tools 2 years on First test - photo-realism Second test - following instructions Third test - being creative My overall verdict on the results of the tests (this blog)

Some other pages relevant to these blogs include:

AI (Artificial Intelligence) classes

You can also book hourly online consultancy for your time zone with one of our 7 expert trainers!

This blog has 0 threads

Add a new post

ALL BLOGS

ARTIFICIAL INTELLIGENCE BLOGS

ARTIFICIAL INTELLIGENCE (AI) BLOGS

In 4th place

In 3rd place

Joint first place

Head office

London

Manchester