Images
Image models from $0.009 an image
10 image models from Black Forest Labs, ByteDance, Alibaba, Google, OpenAI and Krea, behind the same key you use for chat. One request format, one bill, and you see the exact cost of every image.
The OpenAI images format
curl https://api.sparrowcodes.com/v1/images/generations \
-H "Authorization: Bearer $SPARROW_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model": "black-forest-labs/flux-3-klein",
"prompt": "a coffee mug that says ship it",
"aspect_ratio": "1:1", "n": 1}'Images come back as base64 in data[].b64_json, and the response includes its cost in US dollars.
Same prompt, every model
See the difference before you pay for it
One prompt, one image from each model, default settings. The prompt asks for a mug that says “ship it”.
Prices
Every image model
Prices are per image at the default size. Models billed by tokens show a typical price; the real cost depends on size and quality and is returned with every response.
| Model | Options | Per image |
|---|---|---|
| Recraft V4.1 Flash recraft/recraft-v4.1-flash | 5 aspect ratios · up to 4 per request | $0.009 |
| FLUX 3 Klein black-forest-labs/flux-3-klein | 7 aspect ratios · 1K or 2K · up to 4 per request | $0.011 |
| Krea 2 Medium Turbo krea/krea-2-medium-turbo | 1024x1024 · up to 4 per request | $0.018 |
| Seedream 5.0 Flash bytedance-seed/seedream-5-flash | 3 sizes · up to 4 per request | $0.021 |
| Qwen Image 3 qwen/qwen-image-3 | 3 sizes · up to 4 per request | $0.035 |
| Nano Banana 2.1 google/nano-banana-2.1 | 1024x1024 · up to 4 per request | $0.041 |
| FLUX 3 black-forest-labs/flux-3 | 7 aspect ratios · 1K or 2K · up to 4 per request | $0.056 |
| GPT Image 2 openai/gpt-image-2 | 3 sizes · 3 quality levels · up to 10 per request | from $0.30 |
| GPT Image 2.6 Sunburst openai/gpt-image-2.6-sunburst | 3 sizes · 3 quality levels · up to 10 per request | from $0.30 |
| GPT Image 2.6 Flare openai/gpt-image-2.6-flare | 3 sizes · 3 quality levels · up to 10 per request | from $0.30 |
Why here
Image generation without another account
- One key for everything
- The same key, balance and spend limits as your chat and speech requests.
- Switch models by name
- Change
modelto compare results. Size, aspect ratio and quality options are checked before you're charged. - Pay per image
- No subscription for images. Failed generations are not charged.
- Text in images
- Seedream, Qwen Image and FLUX 3 render short text well, including Chinese with Qwen.
- Several at once
- Ask for variations in one request with
n: up to 4 on most models, 10 on GPT Image. - Every request logged
- Model, cost and time for every image in your activity log, exportable to CSV.
Generate your first image
Sign in with your email, add credits, and call any image model with the key you already have.





