local/flux-2-klein-9b Benchmark — 57/100 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.

MakerBlack Forest Labs

FamilyFLUX.2

Model size9B9B params

Costself-hostedfree

Latency8.5s median

Run targetlocal/flux-2-klein-9b

Effective requestmodel: flux-2-klein-9b · size: 1024x1024 · seed: 42

SourcesModel cardWeights & checkpoint

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

63%

Typography Style33%

Writing accuracy42%

Spatial Reasoning

58%

Attributes Binding56%

Compositionality67%

Counting56%

Negation67%

Relative Position58%

Scale & Proportions67%

Human realism

56%

Faces & Expressions42%

Full Body50%

Hands58%

Multi-Subject50%

Professional Studio

60%

Camera & Lighting58%

Color Precision42%

Photorealism67%

Graphical design

44%

Data Visualisation33%

Layout & Design44%

Style Diversity33%

Truthfulness

37%

Photorealism33%

Physics & Reflections25%

World Knowledge25%