Outfit Transfer Challenge

Vote
Image Editing Photorealism

5 models were given the same image and edit instruction, and the community voted blind on which outputs looked best. How it works

This is extremely practical (think e-commerce, fashion, virtual dressing rooms). It’s visually very obvious when it fails, and success looks impressive. It tests a model’s understanding of clothing physics, body shape, and lighting.

Edit instruction
Use Image 1 as the base person. Dress them in the exact elaborate outfit from Image 2 (including all layers, accessories, jewelry, and shoes). Carefully adapt the clothing to the body shape and pose in Image 1 while maintaining realistic fabric behavior, correct proportions, and perfect lighting/shadow matching. Keep the person’s exact face, hair, and background completely unchanged.
Voters were asked to judge by Clothing Accuracy Fit & Fabric Realism Lighting & Shadow Consistency No Side Effects

Challenge Rankings

5 models
# Model Elo
1 1095
2 1082
3 1062
4 1059
5 996

Alibaba’s Wan 2.7 leads the leaderboard with an Elo of 1095 and a 62.5% win rate, narrowly outpacing OpenAI's GPT Image 1 Mini. Notably, the budget-friendly GPT model (1082 Elo) delivers a similar 62.5% win rate at one-sixth the price per image compared to Wan 2.7.

Elo vs Speed

3 models waiting for enough speed data

Competitors

5 models, ranked by Elo
1

Wan 2.7

Try in Playground →
2

GPT Image 1 Mini

Try in Playground →
3

Qwen Image Edit Latest

Playground coming soon

FLUX.2 [klein] 9B

Try in Playground →

Grok Imagine Image

Try in Playground →