AI Potluck
Back to Gap Map Model components / Image, video, 3D & music generation

Midjourney

Midjourney
closed / Overall score: n/a

Midjourney is an independent research lab's image generation service, used through its website and Discord. Its models generate images from text and image prompts with style and mood references, and the service adds image editing and an image-to-video mode that animates still images into short clips.

The arena lists Midjourney's V6 and V7 models; its newer V8 models had no arena entry when read.

Openness

1 high confidence
1.0
weights
closed(no weights are distributed
data
closed
code
closed
license
Proprietary(Midjourney terms of service

Midjourney runs only as a hosted service on its website and Discord. No weights, training data or code are published.

Adoption

not assessed

Midjourney publishes no user or generation count, and a hosted service has no download count, so no reading was possible.

Capability

2 medium confidence
2.0

The Midjourney models on the text-to-image arena place in its lower half, close to Stable Diffusion 3.5. The arena has not yet ranked Midjourney's current V8 models, which may place higher.

Verified 2026-09-26