Which prompts will free-tier image generation be able to produce by end of 2024?
17
Ṁ2824
Dec 31
94%
A person running with one shoe on
62%
A dog with a hat on his tail, but no hat on his head
41%
A cluttered room with a really small cat in it, so small it's hard to spot it on first sight
16%
Exactly 17 soccer balls

Will image generators become good enough for these prompts?

Each option will resolve to YES when either

  1. I will put the prompt in image generation AI tool and the result will obey the prompt.

  2. Anyone post a public link to such generation he made itself.

Any free publicly available image generation tool is OK, except esoteric ones (custom fine-tuning model or one with less than 1k MAU). Tools that internally alters prompts are fine, as long as it is 100% automated.

No additional text (except "please generate an image of ..." when needed for multimodal models), explicit feedback or reference images allowed.

The quality of the generation will be judged by me, and I may post a poll on unclear cases.

#AI
#Technical AI Timelines
#AI Image Generation
Get Ṁ1,000 play money
6 Comments
Sort by:
A dog with a hat on his tail, but no hat on his head

@0482 GPT4o

When I try to copy link it says, "Failed to copy link to clipboard - could not create link"

Is it the free tier? I can't make him generate any image at all, getting only the claim it can't generate images

To be clear, I'll take your word for it, just making sure it was generated under the conditions of free tier

A person running with one shoe on

@0482 this is what Adobe Firefly generate

A cluttered room with a really small cat in it, so small it's hard to spot it on first sight

@0482 this is what Gemini currently generate

Not so small cat

Exactly 17 soccer balls

@0482 An example to what Gemini currently generate for this prompt

Related questions