ImageBench
Model comparison

GPT Image 2 vs Flux 2 vs Nano Banana 2

A side-by-side benchmark comparison of the strongest ImageBench V1 models, including overall pass rate, category strengths, and generated-image evidence.

openai/gpt-image-2 generated image comparison example
openai/gpt-image-2
96.4% overall
fal/google/nano-banana-2 generated image comparison example
fal/google/nano-banana-2
95.3% overall
bfl/flux-2-max generated image comparison example
bfl/flux-2-max
91.7% overall
bfl/flux-2-pro generated image comparison example
bfl/flux-2-pro
83.3% overall

Overall results

ModelOverallPass / FailAvg latencyBenchmarkText RenderingSpatial ReasoningHuman realismTruthfulnessProfessional StudioGraphical design
openai/gpt-image-296.4%185 / 745.3sView result 100.0%96.5%97.6%92.6%92.6%100.0%
fal/google/nano-banana-295.3%183 / 928.1sView result 86.7%96.5%97.6%92.6%100.0%91.7%
bfl/flux-2-max91.7%176 / 1626.7sView result 100.0%93.0%100.0%88.9%88.9%75.0%
bfl/flux-2-pro83.3%160 / 3211.8sView result 80.0%87.7%85.7%70.4%100.0%66.7%

Benchmark result pages

Open the full ImageBench V1 result page for each model, including category breakdowns, prompts, generated images, and pass/fail judgments.

Category takeaways

Text Rendering

openai/gpt-image-2 leads this category at 100.0% (15/15 pass).

openai/gpt-image-2 generated image for Text Rendering

A poster with headline 'SUMMER SALE' in large bold text, body text 'Up to 50% off all items this weekend only' in medium size, and fine print 'Terms and conditions apply. See store for details.' at the bottom in small text

Spatial Reasoning

openai/gpt-image-2 leads this category at 96.5% (55/57 pass).

openai/gpt-image-2 generated image for Spatial Reasoning

A pink fire truck, a blue banana, and a green sun in a surreal landscape

Human realism

bfl/flux-2-max leads this category at 100.0% (42/42 pass).

bfl/flux-2-max generated image for Human realism

An elderly woman in her 80s with deep wrinkles and white hair, smiling warmly with kind eyes

Truthfulness

openai/gpt-image-2 leads this category at 92.6% (25/27 pass).

openai/gpt-image-2 generated image for Truthfulness

A row of trees reflected in a calm lake at sunset, with the reflection slightly darker and gently rippled

Professional Studio

fal/google/nano-banana-2 leads this category at 100.0% (27/27 pass).

fal/google/nano-banana-2 generated image for Professional Studio

A portrait shot at 85mm f/1.4 with creamy bokeh background, backlit with warm rim light separating the subject from a blurred garden

Graphical design

openai/gpt-image-2 leads this category at 100.0% (24/24 pass).

openai/gpt-image-2 generated image for Graphical design

A magazine double-page spread with a left page containing a large hero photo, and a right page with a two-column article, a pull quote, a sidebar with author bio, and a page number