What is the best AI text-to-image model?

Based on blind community voting, GPT Image 2 is currently the #1 ranked AI text-to-image model with an Elo rating of 1301. Rankings update in real time as new votes come in.

How are AI text-to-image 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.

Best AI Models for Art

Art · Text-to-Image

Elo rankings from blind votes across 2 challenges in this category.

Imagen 4.0 Generate 001

Imagen 4.0 Ultra Generate 001

Highlights

10 images

Price · Performance

Best models, by

Every ranked model plotted by price against its Elo. The upper-left frontier is the best value: top quality at the lowest cost.

Best Text-to-Image Models by Price

1 model without pricing omitted.

Best AI Models for Art

56 models ranked · Last update: October 1, 2026 4:16 AM

#	Model	Price	¢/img	Elo
1	GPT Image 2 OpenAI	$$	3.9¢	1301
2	Qwen Image 2512 Alibaba	$$	2¢	1284
3	GPT Image 1.5 OpenAI	$	0.9¢	1279
4	Imagen 4.0 Generate 001 Google	$$	4¢	1243
5	DALL-E 3 OpenAI	$$	4¢	1229
6	Recraft V4 Pro Recraft AI	$$$$	25¢	1227
7	FLUX.2 [klein] 9B Black Forest Labs	$	0.6¢	1226
8	DALL-E 2 OpenAI	$$	1.6¢	1223
9	Qwen Image Alibaba	$$	2¢	1221
10	HiDream I1 Fast HiDream AI	$	0.5¢	1220

11	Imagen 4.0 Ultra Generate 001 Google	$$$	6¢	1213
12	GPT Image 1 Mini OpenAI	$	0.5¢	1208
13	Qwen Image 2.0 Alibaba	$$	3.5¢	1208
14	Wan 2.5 (Preview) Alibaba	$$$	5¢	1206
15	Seedream 4.5 ByteDance	$$	4¢	1206
16	GPT Image 1 OpenAI	$$	1.1¢	1205
17	Recraft V4.1 Recraft AI	$$	4¢	1202
18	Recraft V4 Recraft AI	$$	4¢	1201
19	FLUX.2 [dev] Turbo fal	$	0.8¢	1195
20	Nano Banana Pro Google	$$$	6.7¢	1195
21	FLUX.2 [klein] 4B Black Forest Labs	$	0.1¢	1193
22	FLUX.2 [pro] Black Forest Labs	$$	1.5¢	1191
23	Imagen 3.0 Generate 002 Google	$$	3¢	1190
24	Wan 2.7 Pro Alibaba	$$$	7.5¢	1189
25	Z-Image Turbo Alibaba	$	0.5¢	1179
26	FLUX.2 [flex] Black Forest Labs	$$$	6¢	1178
27	FLUX1.1 [pro] Ultra Black Forest Labs	$$$	6¢	1177
28	Nano Banana 2 Google	$$	2.2¢	1175
29	FLUX.1 Kontext [max] Black Forest Labs	$$$	8¢	1175
30	FLUX.2 [dev] Black Forest Labs	$$	1.2¢	1173

GPT Image 2 leads the Art category with a dominant 91.2% win rate and 1301 Elo, maintaining its position as the overall leaderboard champion. Qwen Image 2512 follows in second place at 1284 Elo, offering a 48% lower price point and significantly faster generation speeds than the top-ranked OpenAI model.

Think a model is ranked wrong? Cast your vote

Highlighted challenges

GPT Image 2 is first to make it GPT Image 2

Text-to-Image Art Photorealism

The Reversed Rodeo

This competition tests how well AI image models truly understand language versus how much they rely on visual habits from their training data. The prompt is deliberately simple on the surface but devilishly hard in practice. Most models default to the familiar trope of an astronaut riding a horse. By forcing the reversal, we measure three critical capabilities that separate good models from great ones: Strict instruction following (including negations) Accurate subject-object relationships and spatial hierarchy Resistance to strong dataset biases

Text-to-Image 3D Imaging & Modeling Photorealism

Geometric Composition

A spatial-reasoning test. Each object has a precise relationship (inside, on top, behind, seen through the glass), so it measures whether a model follows explicit placement instructions and handles transparency and refraction rather than just approximating the scene.