By skyforbes 
Quick context. I shifted from ChatGPT to Grok several months ago. I found that Grok was more succinct with responses and it seemed to use more current information. I am a developer and I found that Grok was consistent in its coding related answers. I have been happy with Grok. I’m not saying Grok is the best, I’m saying it has been working for me. But not for image creation.
I wasn’t trying to test all image creation tools. I wasn’t trying to be perfect with my prompts. I was simply trying to create an image that mocked the Monthly Challenge in Apple Watch. I know that there are more precise methods to accomplish this. I’m simply sharing my experience.
Scenario.
I was working on an article related to Apple Watch and closing circles in the Fitness app. Like many, I find the Monthly Challenges to be unwieldy. They’re either incredibly easy or they go the opposite direction and give you a nearly impossible goal.
In support of the article, I wanted an image of the Monthly Challenge with an obvious impossible goal.
Full disclosure, I pay for SuperGrok and use the free version of ChatGPT.
In Grok, I used the Imagine feature. This resulted in a lot of images and it did so very fast. Unfortunately, they weren’t even close to what I was looking for.
Time to change my approach. I uploaded an image that was taken from my Apple Watch.
At this point, Imagine created a video that still missed the mark. I understand that some of this may be my unfamiliarity with Imagine. Nonetheless, I didn’t get what I was looking for.
I kept refining, but the closest I got was this. What is this?
Time to try this on ChatGPT.
Using ChatGPT, I prompted in the same manner. NOTE: This was a slow process for the first generation. Open AI was quick to tell me that I could pay for Plus and the creation would be much faster.
First prompt
this is an image from my Apple Watch. I'd like a new image and it should look nearly identical, but the goal should be unattainable
This resulted in an ok image
Updated prompt
Please regenerate this, but keep the title as November Challenge. Show the detail as 100 miles per day. Show that I haven't completed it at all. The progress bar should be blank.
This resulted in a usable image
This made me wonder, what is the “best” image generation tool? A simple Google search (remember Google?) resulted in the following AI Overview:
For the best overall image generation, GPT-4o is the strongest option due to its ability to generate highly accurate and detailed images directly within the ChatGPT interface, including legible text. However, the “best” choice depends on your needs
Fair enough.
Keep in mind, I don’t have a preference. I’m using these models as tools to get things done. I know that some of you are going to tell me that I should use Midjourney or Stable Diffusion or [insert model name here].
AI isn’t magical, but it’s pretty amazing. Different tools for different purposes. I am a Grok user, but I’m always willing and looking to expand my toolset.