ImageBench

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 ImageBench V1 run.

replicate/luma/photon-flash

API

Photon Flash is the speed-optimized variant of Luma AI's Photon model, built for fast, low-cost image generation while retaining much of Photon's visual quality. It supports character consistency and multi-image references, positioned for high-volume use.

Maker
Luma AI
Family
Photon
Model Size
not disclosed
unknown
Cost
$0.01 / image
high
Run Target
replicate/luma/photon-flash
Effective Request
aspect_ratio: 1:1 · seed: 7
32.9
Overall
46%
Capability
20.0
Est. Preference
88
Pass
104
Fail
6.4s
Avg Latency
4.0s
Min Latency
11.0s
Max Latency
Text Rendering47%Spatial Reasoning42%Human realism38%Truthfulness33%Professional Studio85%Graphical design38%Preference20%Latency45%

All 192 generations

Text Rendering47%

Typography Style100%

Writing accuracy33%

Spatial Reasoning42%

Attributes Binding67%

Compositionality78%

Counting11%

Negation22%

Relative Position33%

Scale & Proportions44%

Human realism38%

Faces & Expressions67%

Full Body8%

Hands50%

Multi-Subject17%

Truthfulness33%

Photorealism0%

Physics & Reflections33%

World Knowledge42%

Professional Studio85%

Camera & Lighting75%

Color Precision100%

Photorealism67%

Graphical design38%

Data Visualisation0%

Layout & Design0%

Style Diversity75%