Datasets:
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
dataset_info:
features:
- name: prompt
dtype: string
- name: image1
dtype: image
- name: image2
dtype: image
- name: model1
dtype: string
- name: model2
dtype: string
- name: weighted_results_image1_preference
dtype: float32
- name: weighted_results_image2_preference
dtype: float32
- name: detailed_results_preference
dtype: string
- name: weighted_results_image1_coherence
dtype: float32
- name: weighted_results_image2_coherence
dtype: float32
- name: detailed_results_coherence
dtype: string
- name: weighted_results_image1_alignment
dtype: float32
- name: weighted_results_image2_alignment
dtype: float32
- name: detailed_results_alignment
dtype: string
splits:
- name: train
num_bytes: 34055275649
num_examples: 44857
download_size: 28906555375
dataset_size: 34055275649
license: cdla-permissive-2.0
task_categories:
- text-to-image
- image-to-text
- image-classification
- reinforcement-learning
language:
- en
tags:
- Human
- Preference
- Coherence
- Alignment
- country
- language
- flux
- midjourney
- dalle3
- stabeldiffusion
- alignment
- flux1.1
- flux1
- imagen3
- aurora
- lumina
- recraft
- recraft v2
- ideogram
- frames
- OpenAI 4o
- 4o
- OpenAI
- Seedream-3
- seedream
- Imagen-4
- Google
- Recraft v3
- Hunyuan Image 2.1
- Flux 2 Pro
size_categories:
- 10K<n<100K
pretty_name: >-
Flux 2 Pro vs Hunyuan Image 2.1 / Recraft v3 / 4 Ultra 24.7.25 / Seedream 3 /
Ideogram V2 / Recraft V2 / Lumina-15-2-25 / Frames-23-1-25 / Aurora / imagen-3
/ Flux-1.1-pro / Flux-1-pro / Dalle-3 / Midjourney-5.2 / Stabel-Diffusion-3 /
4o - Human Preference Dataset
Rapidata Flux 2 Pro Preference
This T2I dataset contains over ~400'000 human responses from over ~50'000 individual annotators, collected in less than 7h using the Rapidata Python API, accessible to anyone and ideal for large scale evaluation. Evaluating Flux 2 Pro (version from 25.11.25) across three categories: preference, coherence, and alignment.
Explore our latest model rankings on our website.
If you get value from this dataset and would like to see more in the future, please consider liking it ❤️
To add your own model to the benchmark send us an e-mail at: jason@rapidata.ai
Overview
The evaluation consists of 1v1 comparisons between Flux 2 Pro (version from 24.7.2025) and 18 other models:
- 4o
- Flux-1-pro
- Flux-1.1-pro
- imagen 4 ultra
- Aurora
- Imagen-3
- DALL-E 3
- Midjourney-5.2
- Frames-23-1-25
- Stable Diffusion 3
- Janus-7b.
- hidream-l1-full
- Recraft V2
- Ideogram V2
- halfmoon-4-4-25
- Lumina-15-2-25
- Imagen 4 Ultra 20.5.25
- Imagen 4 Ultra 24.7.25
- Recraft v3
- Hunyuan Image 2.1
Alignment
The alignment score quantifies how well an video matches its prompt. Users were asked: "Which image matches the description better?".
The cracked rectangle was leaning against the glossy cylinder and the peeling triangle.
Flux 2 Pro
Seedream 3
The square coaster was next to the circular glass.
Flux 2 Pro
4o
Coherence
The coherence score measures whether the generated video is logically consistent and free from artifacts or visual glitches. Without seeing the original prompt, users were asked: "Which image has more glitches and is more likely to be AI generated?"
Flux 2 Pro
Flux 1 Pro
Flux 2 Pro
Halfmoon 4
Preference
The preference score reflects how visually appealing participants found each image, independent of the prompt. Users were asked: "Which image do you prefer?"
Flux 2 Pro
Stable Diffusion 3
Flux 2 Pro
Lumina
Benchmark
About Rapidata
Rapidata's technology makes collecting human feedback at scale faster and more accessible than ever before. Visit rapidata.ai to learn more about how we're revolutionizing human feedback collection for AI development.