Skip to content

AI Image Generation Cost Analysis

9 min

I've benchmarked 31 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.

Key findings (November 6, 2026): GPT Image 2 had the lowest recorded cost at $0.00615 per image, and Flux Pro 1.1 was the fastest at 3.8 seconds. These results come from one prompt with model-specific settings, so they don't establish the best model for every task.

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

Full Results

November 6, 2026: added Flux 3, Flux 3 Klein, Nano Banana 2.1, MAI-Image-2.6, MAI-Image-2.6 Flash, GPT Image 2.6 Flare, and Recraft V4.1 Flash; removed GPT Image 1 after it disappeared from the live AI Gateway catalog and Flux Schnell after Prodia retired it; and dropped seven models superseded by cheaper and faster versions (GPT Image 1.5, GPT Image 1 Mini, Recraft V3, Recraft V4, Recraft V4 Pro, Flux Kontext Max, and Nano Banana Classic). All results, thumbnails, and charts were regenerated. Full changelog

(click on a thumbnail to see the full image)

Model Price Latency Image
google/gemini-3.1-flash-image $0.067024 10.6s Nano Banana 2
google/gemini-nano-banana-2.1 $0.040332 22.0s Nano Banana 2.1
google/gemini-3.1-flash-lite-image $0.034012 12.8s Nano Banana 2 Lite
google/gemini-3-pro-image $0.13746 31.9s Nano Banana Pro
bfl/flux-2-flex $0.05 12.3s Flux 2 Flex
bfl/flux-2-klein-4b $0.014 4.8s Flux 2 Klein 4B
bfl/flux-2-klein-9b $0.015 5.3s Flux 2 Klein 9B
bfl/flux-2-pro $0.03 10.4s Flux 2 Pro
bfl/flux-2-max $0.07 19.7s Flux 2 Max
bfl/flux-3-image $0.048 28.8s Flux 3
bfl/flux-3-image-klein $0.009 7.0s Flux 3 Klein
bfl/flux-pro-1.1 $0.04 3.8s Flux Pro 1.1
bfl/flux-pro-1.1-ultra $0.06 7.4s Flux Pro 1.1 Ultra
bfl/flux-kontext-pro $0.04 6.0s Flux Kontext Pro
recraft/recraft-v2 $0.022 6.6s Recraft V2
recraft/recraft-v4.1 $0.035 7.6s Recraft V4.1
recraft/recraft-v4.1-flash $0.007 6.0s Recraft V4.1 Flash
recraft/recraft-v4.1-pro $0.21 19.6s Recraft V4.1 Pro (2048x2048)
recraft/recraft-v4.1-utility $0.035 10.6s Recraft V4.1 Utility
recraft/recraft-v4.1-utility-pro $0.21 23.1s Recraft V4.1 Utility Pro (2048x2048)
openai/gpt-image-2 $0.00615 13.2s GPT Image 2
openai/gpt-image-2.6-flare $0.02499 10.9s GPT Image 2.6 Flare
meta/muse-image-1.0 $0.01 14.5s Muse Image 1.0
microsoft/mai-image-2.6 $0.039147 27.0s MAI-Image-2.6
microsoft/mai-image-2.6-flash $0.01953825 13.4s MAI-Image-2.6 Flash
spacexai/grok-imagine-image $0.02 7.8s Grok Imagine
spacexai/grok-imagine-image-2.0 $0.04 16.6s Grok Imagine Image 2.0
bytedance/seedream-4.0 $0.03 11.1s Seedream 4.0 (2048x2048)
bytedance/seedream-4.5 $0.04 11.1s Seedream 4.5 (2048x2048)
bytedance/seedream-5.0-lite $0.035 27.5s Seedream 5.0 Lite (2048x2048)
bytedance/seedream-5.0-pro $0.035 44.5s Seedream 5.0 Pro (2048x2048)

Total spent: $1.47365325

Each model generated one image, so each latency is a single measurement.
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

AI image generation cost chart

Latency

AI image generation latency chart

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-v4.1-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

Winners

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

  • openai/gpt-image-2 at $0.00615
  • recraft/recraft-v4.1-flash at $0.007
  • bfl/flux-3-image-klein at $0.009

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

  • bfl/flux-pro-1.1 at 3.8s
  • bfl/flux-2-klein-4b at 4.8s
  • bfl/flux-2-klein-9b at 5.3s

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.

I've also run a similar benchmark for 18 text-to-video models through the same gateway.

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.

Relevant Posts