Generation highlights
How the top models compare
Quality bought per dollar
Best Image Upscaling Models by Price
1 model without pricing omitted.
Best AI Models for Image Upscaling
| # | Model | Elo |
|---|---|---|
| 1 | 1288 | |
| 2 | 1180 | |
| 3 | 1155 | |
| 4 | 1132 |
As of July 2026, Crystal Upscaler dominates the leaderboard with a 1291 Elo and a substantial 95.9% win rate, maintaining a 116-point gap over its nearest rival, Recraft Crisp Upscale (1175 Elo). While the top spot is uncontested, the competition for second remains tight as Google Upscaler follows Recraft within a narrow 20-point margin. Notably, the budget-tier Recraft Crisp Upscale achieves the #2 position at just 25% of the leader's cost per image.
Rivalries
Aggregate head-to-head across the arena
Highlighted challenges
Weave Warfare
This challenge is highly revealing for an arena as it directly exposes whether models can maintain complex woven textures without introducing unnatural pattern repetition or losing the authentic fabric feel.
Freckle Fortress Challenge
This challenge is excellent for an arena because it forces AI models to prove they can enhance ultra-fine organic skin details without destroying freckles or turning skin into an unnatural plastic or overly smooth surface.
Awaiting blind votes
2 models have taken it on, but there aren't enough votes to rank them yet. Open the challenge to see every take.
Recent SOTA shifts in Image Upscaling
Full historyFAQ
What is the best AI image upscaling model?
Based on blind community voting, Crystal Upscaler is currently the #1 ranked AI image upscaling model with an Elo rating of 1291. Rankings update in real time as new votes come in.
How are AI image upscaling models ranked on Lumenfall?
Lumenfall Arena ranks AI models through blind community voting. In each matchup, two models generate from the same prompt and voters pick the better result without seeing model names. Votes are processed using TrueSkill, a Bayesian rating algorithm developed by Microsoft Research, that produces a single Elo score reflecting each model's relative quality.
What is an Elo rating for AI models?
An Elo rating is a numerical score representing a model's skill relative to other models. Under the hood, Lumenfall uses TrueSkill, which tracks two values per model: mu (estimated skill) and sigma (uncertainty). The displayed Elo is calculated as 1000 + 10 x (mu - 3*sigma), a conservative lower bound. A model must prove itself consistently across many matchups to earn a high rating.
Keep the arena honest
Cast your vote
Pick winners in blind matchups. Every vote nudges the Elo and shapes these rankings.
Cast Your VoteSuggest a prompt
Got an idea worth testing? Submit a prompt and watch the models battle it out.