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.

fal/google/nano-banana-2

API

Nano Banana 2 is Google's fast Gemini image generation and editing model, exposed here through fal, with emphasis on high-quality image creation, grounded world knowledge, and practical generation speed.

Maker: Google DeepMind
Family: Nano Banana / Gemini image generation
Model Size: not disclosed
unknown
Cost: $0.08 / image at 1K
high
Run Target: fal/fal-ai/nano-banana-2
Effective Request: aspect_ratio: 1:1 · resolution: 1K · num_images: 1 · output_format: png
Sources: Google Gemini image generation overview Google Cloud Nano Banana 2 and Nano Banana Pro GA Nano Banana 2 on fal

73.0

Overall

73%

Capability

72.6

Est. Preference

141

Pass

51

Fail

28.1s

Avg Latency

13.8s

Min Latency

135.4s

Max Latency

All 192 generations

Text Rendering73%

Typography Style100%

Writing accuracy67%

Spatial Reasoning88%

Attributes Binding100%

Compositionality100%

Counting78%

Negation89%

Relative Position92%

Scale & Proportions67%

Human realism38%

Faces & Expressions92%

Full Body8%

Hands25%

Multi-Subject17%

Truthfulness56%

Photorealism67%

Physics & Reflections58%

World Knowledge50%

Professional Studio93%

Camera & Lighting92%

Color Precision100%

Photorealism67%

Graphical design100%

Data Visualisation100%

Layout & Design100%

Style Diversity100%