Resolves as YES if the best available text to image model from Midjourney beats the best available model from OpenAI in January 2025 for this prompt:
"A seagull firing a longbow while riding a tricycle next to a five-legged dog wearing jeans"
The test will occur on the first days of January 2025, and will take into consideration the 10 first images generated by the best available model from each company.
Only text to image with be considered (no inpainting or other guidance techniques).
As of December 11th 2023, here are the current best images:
(Generated using Dalle-3 / Bing image creator)
(Generated using Midjourney 5.2)
@0482 consistency with the prompt first, and then image quality / aesthetics second.
I will run this and post the images here, rating each image pair (one from either model)