MethodologyAnnotator qualityUpdatesHugging Face438.2k downloadsSuggest a model or benchmarkRequest a private benchmark
Benchmarks
Dimensions
License
Models
Categories
Capability
Output Type
Subject
View
Text-to-Image v1updated 2026-07-23

The Human-Preference Leaderboard for Text-to-Image Generation

Rapidatabenchmark@rapidata.ai

Abstract

This benchmark ranks text-to-image models using human comparisons of images generated from the same prompt. We curated 1,514 prompts across 25 categories, focusing on commercial applicability and ensuring reproducibility of the results. The rankings are updated continuously from 7,844,896 responses across 1,569,555 pairwise human-preference matches. All data and judgments are public; see the methodology for details on evaluation and scoring.

LeaderboardCompare OutputsMatchupsAnalysisPromptsDetails
#
Model
Organization
Score
95% CI
Interval
Responses
Matches
Win rate
#
Model
Score

Table 1. Overall standings, General Image benchmark. Scores from a Bradley–Terry fit rescaled to Elo; intervals are 95% Wald intervals. Click any row to view sample outputs in Figure 1.

Methodology notesCompare Outputs
Figure 1. Sample outputs from GPT Image 2.6 (Sunburst) on representative prompts from the fixed set. Outputs shown are those used in matches.
Score vs. cost
80010001200$0.005$0.01$0.02$0.05$0.1$0.2cost, USD per image (log scale)score

Figure 2. Overall Elo of each model against its list price per image, on a logarithmic axis. Prices are recorded manually from the providers’ published rates. Error bars are 95% confidence intervals of the score. Models ringed in teal and joined by the step line form the Pareto frontier: no other model has both a lower cost and a higher score. The remaining models are faded. 3 models without a cost in USD per image are not shown.

Evaluation: Models are compared on the same prompts through pairwise human judgements.
Score: Scores show relative performance within this benchmark. A higher score means a model was more likely to be preferred over the other evaluated models.
Uncertainty: The 95% confidence interval shows the uncertainty around each model’s estimated score.
Views: Results can be explored by overall preference, coherence, prompt alignment, and by prompt category.
Coverage: This benchmark includes a fixed prompt set across multiple categories, evaluated on all ranked models.
Data: Available prompts, model outputs, and evaluation results can be downloaded for further analysis.
Version: Results correspond to Text-to-Image v1, last updated on 2026-11-20.

We use crowd intelligence to deliver real, diverse human feedback at scale for model evaluation and post-training.

Benchmarks
Text-to-ImageOutput Diversity (Mode Collapse)Multi-Step Image Editing (Identity Drift)Camera MovementRelative Camera MovementPhysicsPhysicsRealismImage-to-3DSVGLLM-to-3DLLM to AnimationSuggest a model or benchmark
Method
MethodologyData on Hugging Face
Rapidata
rapidata.ai Private benchmarksContactX LinkedIn
Rapidata, benchmark.ai © 2026