AI Image Generation Cost Analysis

I've benchmarked 18 image generation models to compare cost and latency using the same prompt:

A photorealistic image of a golden retriever puppy sitting in a sunflower field at golden hour, with a soft bokeh background and warm light. Aspect ratio: 1:1, size: 1024x1024.

The benchmark includes both image-only and image-capable multimodal models available on Vercel AI Gateway.

If you don't have much time to analyze the full report:

• go straight to the winners
• compare images
• check out the cost chart
• or the latency chart

Full Results

Updated on April 17, 2026

Click on a thumbnail to see the full image.

Model	Price	Latency	Image
google/gemini-2.5-flash-image	$0.04*	7.1s
google/gemini-3.1-flash-image-preview	$0.07*	34.6s
google/gemini-3-pro-image	$0.14*	19.4s
google/imagen-4.0-fast-generate-001	$0.02	5.6s
google/imagen-4.0-generate-001	$0.04	10.3s
google/imagen-4.0-ultra-generate-001	$0.06	13.9s
bfl/flux-2-flex	$0.05	12.6s
bfl/flux-2-pro	$0.03	7.6s
bfl/flux-pro-1.1	$0.04	4.5s
bfl/flux-pro-1.1-ultra	$0.06	8.3s
bfl/flux-kontext-pro	$0.04	6.3s
bfl/flux-kontext-max	$0.08	7.4s
recraft/recraft-v3	$0.04	6.3s
recraft/recraft-v4	$0.04	14.6s
recraft/recraft-v4-pro	$0.25	28.9s	(2048x2048)
openai/gpt-image-1	$0.04*	20.5s
xai/grok-imagine-image	$0.02	3.5s
xai/grok-imagine-image-pro	$0.07	18.3s

* rounded to 2 decimal places

The latency here is wall time, measured by the benchmark script.

The cost, however, is returned by the gateway, so it should be accurate.

Cost

Latency

Model Quirks

This benchmark includes provider-specific adaptations, because several models need custom params to work correctly:

• Black Forest Labs Flux models need explicit pixel width and height instead of relying on aspectRatio
• Recraft and OpenAI image models require size rather than aspectRatio
• recraft/recraft-v4-pro needs 2048x2048 for square output, not 1024x1024
• xAI Grok models support aspectRatio but not size
• Gemini image responses can contain duplicate files in result.files, so the benchmark deduplicates them before saving
• xai/grok-imagine-image-pro failed when run last in a long benchmark sequence, so isolated runs may be more reliable for it

Winners

If you only optimize for cost, the two cheapest models in this benchmark are:

• google/imagen-4.0-fast-generate-001 at $0.02
• xai/grok-imagine-image at $0.02

If you optimize for latency, the two fastest models are:

• xai/grok-imagine-image at 3.5s
• bfl/flux-pro-1.1 at 4.5s

I personally think it all depends on your task and prompt. For my particular mass-scale case, I'm optimizing for both cost and speed, so I'll probably choose among these three models.

DIY

If you want to run the benchmark yourself, add more models, or adjust the prompt - the script is easy to configure and run: kometolabs/ai-image-generation-cost-analysis