fal/krea/v2-large — ImageBench V1
ImageBench V1 —
192 evaluations across 6 categoriesBenchmark V1 verdicts are produced by VLM judges and can contain mistakes. Treat PASS/FAIL labels as machine-assisted assessments, and inspect the images yourself. Learn more about the methodology.
Generation Details
Source-backed model context, size, cost, and request settings for this run.
MakerKrea
FamilyKrea
Model sizenot disclosedundisclosed
Cost$0.060 / imagemid
Latency30.1s median
Run targetfal/krea/v2-large
Effective requestmodel: v2-large · size: 1024x1024 · seed: 42
SourcesKrea model docsKrea pricing
All 192 generations
Every challenge, grouped by category and subcategory. Solid-bordered tiles with a check passed the VLM judge; dashed tiles with a cross failed. Images are placeholders until generated.
Text Rendering
44%Typography Style33%
Writing accuracy25%
Spatial Reasoning
71%Attributes Binding67%
Compositionality56%
Counting56%
Negation67%
Relative Position42%
Scale & Proportions78%
Human realism
69%Faces & Expressions50%
Full Body75%
Hands50%
Multi-Subject83%
Professional Studio
54%Camera & Lighting42%
Color Precision33%
Photorealism67%
Graphical design
71%Data Visualisation67%
Layout & Design78%
Style Diversity58%
Truthfulness
51%Photorealism67%
Physics & Reflections33%
World Knowledge50%