fal/krea/v2-medium-turbo Benchmark — 49/100 Overall, 192 Prompts, Every Image Shown

ImageBench V1 —

192 evaluations across 6 categories

Benchmark 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

Costnot foundvariable

Latency15.5s median

Run targetfal/krea/v2-medium-turbo

Effective requestmodel: v2-medium-turbo · 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

57%

Typography Style67%

Writing accuracy33%

Spatial Reasoning

60%

Attributes Binding56%

Compositionality56%

Counting44%

Negation78%

Relative Position58%

Scale & Proportions67%

Human realism

66%

Faces & Expressions58%

Full Body83%

Hands83%

Multi-Subject50%

Professional Studio

48%

Camera & Lighting42%

Color Precision50%

Photorealism67%

Graphical design

62%

Data Visualisation100%

Layout & Design67%

Style Diversity58%

Truthfulness

49%

Photorealism67%

Physics & Reflections33%

World Knowledge33%