I spent a great deal of time checking the new AI tools such as ChatGPT, Microsoft Copilot and lately X’s Grok-2 (beta) tool.

They are all kind of similar to one another, only difference is the pricing and small details. Of course, their internal algorithms may be running slightly different code, but the output they create is kind of the same. Both Copilot and ChatGPT are using DALL-E engine, while Grok-2 is using its own.

Especially when it comes to creating visuals, all of them have quite a similar take on it. However, personally, I think Microsoft Copilot creates the best visuals. It always gives you a choice of 4 pictures with various tones of the desired output.

Copilot

For example when I tried prompting this code below for Copilot, I get the following results:

prompt: create an image of chihuahua dog wearing Christmas hat, looking strong, smiling, wearing dark sun glasses with Christmas tree in the background, 6 reindeer, and sleigh

Grok-2

Let me try to do the same thing with Grok-2 in X.

The difference is somewhat not noticeable other than it being much more head-shot. But I like the fact that Copilot provides me with 4 different results.

ChatGPT

Now I try ChatGPT to create a visual from exactly the same prompt

ChatGPT image is looking more like Grok-2. The dog is much in the foreground with less noticeable details in the background. All-in-all, Copilot’s visual has a friendlier radiance, with softer colors and more engaging depth to it – so that is my #1 choice for visuals so far!

In any case, whatever algorithms these tools are using, the output is somewhat the same except minor details. And I tried this test with other visual as well, the results are pretty much the same with Copilot being slightly the better version. One other distinctive difference is that Grok-2 is not much restricted and can create images of popular people while ChatGPT and Copilot throw a policy restriction error.

But whatever tool you are using, the key is to use correct prompting. If you master prompting science, then you can get most marvelous results. That is what I will be set on learning in the near future and I will post updates of my learnings.

Cheers!