Pose & Character Mashup
Vote5 models were given the same image and edit instruction, and the community voted blind on which outputs looked best. How it works
This prompt is challenging because it forces models to do two hard things at once: accurately transfer complex body poses and joint angles from Image 1 while strictly preserving the exact face, identity, and clothing style from Image 2. Most models struggle with unnatural limb distortions, identity blending, incorrect fabric behavior on the new pose, and mismatched lighting, making it a strong test of advanced spatial reasoning and character consistency.
#1 — GPT Image 1 Mini
Challenge Rankings
| # | Model | Elo |
|---|---|---|
| 1 | 1179 | |
| 2 | 1134 | |
| 3 | 1100 | |
| 4 | 1079 | |
| 5 | 1051 |
GPT Image 1 Mini leads the leaderboard with an 1179 Elo and a 66.7% win rate, outperforming the high-tier Wan 2.7 Pro by exactly 100 Elo points despite costing 15x less. Budget-friendly models dominate the top of this challenge, with FLUX.2 [klein] 9B (1134 Elo) providing the fastest competitive results at nearly double the speed of the trailing Seedream 5.0 Lite.
Elo vs Cost
Elo vs Speed
Competitors
5 models, ranked by EloHighlighted Battles
The most competitive head-to-head matchups, selected by closeness and vote count.