AI Image Generation Cost Analysis

I've benchmarked 33 image generation models from 8 providers 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

August 28, 2026: added Meta Muse Image 1.0, Seedream 5.0 Pro, and Grok Imagine Image 2.0; migrated the base Grok model to its current spacexai slug; and removed the three Imagen 4 models after they disappeared from the live AI Gateway catalog. All results, thumbnails, and charts were regenerated. Full changelog

(click on a thumbnail to see the full image)

Model	Price	Latency	Image
google/gemini-2.5-flash-image	$0.0390419	15.1s
google/gemini-3.1-flash-image	$0.067024	15.7s
google/gemini-3.1-flash-lite-image	$0.034012	9.9s
google/gemini-3-pro-image	$0.137664	46.0s
bfl/flux-2-flex	$0.05	12.3s
bfl/flux-2-klein-4b	$0.014	3.3s
bfl/flux-2-klein-9b	$0.015	3.6s
bfl/flux-2-pro	$0.03	10.6s
bfl/flux-2-max	$0.07	33.8s
bfl/flux-pro-1.1	$0.04	6.1s
bfl/flux-pro-1.1-ultra	$0.06	9.1s
bfl/flux-kontext-pro	$0.04	6.7s
bfl/flux-kontext-max	$0.08	88.0s
recraft/recraft-v2	$0.022	11.2s
recraft/recraft-v3	$0.04	9.2s
recraft/recraft-v4	$0.04	12.0s
recraft/recraft-v4-pro	$0.25	24.5s	(2048x2048)
recraft/recraft-v4.1	$0.035	8.7s
recraft/recraft-v4.1-pro	$0.21	15.0s	(2048x2048)
recraft/recraft-v4.1-utility	$0.035	12.3s
recraft/recraft-v4.1-utility-pro	$0.21	35.1s	(2048x2048)
openai/gpt-image-1	$0.16667	38.3s
openai/gpt-image-2	$0.00615	18.1s
openai/gpt-image-1.5	$0.144558	37.5s
openai/gpt-image-1-mini	$0.033388	43.4s
meta/muse-image-1.0	$0.01	10.5s
spacexai/grok-imagine-image	$0.02	7.8s
spacexai/grok-imagine-image-2.0	$0.06	73.2s
bytedance/seedream-4.0	$0.03	10.9s	(2048x2048)
bytedance/seedream-4.5	$0.04	9.5s	(2048x2048)
bytedance/seedream-5.0-lite	$0.035	29.3s	(2048x2048)
bytedance/seedream-5.0-pro	$0.035	40.9s	(2048x2048)
prodia/flux-fast-schnell	$0.0025	4.5s

Total spent: $2.1020079

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:

• bfl/flux-pro-1.1 needs explicit pixel width and height to enforce a 1024x1024 output because BFL does not natively support the SDK's aspectRatio parameter
• Recraft and OpenAI image models require size rather than aspectRatio
• Recraft Pro models (recraft-v4-pro, recraft-v4.1-pro, and recraft-v4.1-utility-pro) run at 2048x2048 for square output
• meta/muse-image-1.0 does not support aspectRatio, so it uses the benchmark's configured 1024x1024 size
• SpaceXAI Grok image models support aspectRatio but not size
• Gemini image responses can contain duplicate files in result.files, so the benchmark deduplicates them before saving
• ByteDance Seedream models require size (not aspectRatio) at a minimum of 3.69 MP, so they run at 2048x2048, and take watermark: false to drop the "AI generated" badge they otherwise burn into the image
• prodia/flux-fast-schnell, still the cheapest model, continues to work after its earlier transient provider failure

Winners

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

• prodia/flux-fast-schnell at $0.0025
• openai/gpt-image-2 at $0.00615
• meta/muse-image-1.0 at $0.01

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

• bfl/flux-2-klein-4b at 3.3s
• bfl/flux-2-klein-9b at 3.6s
• prodia/flux-fast-schnell 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'd choose among the winners.

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

Support My Work

Preparing such research sometimes takes 2-5 full benchmark runs, which adds up. If it saved you time, consider sponsoring me on GitHub or Buy Me a Coffee.