falImage to imageAvailable now

openai/gpt-image-2.5/sunburst/edit

openai/gpt-image-2.5/sunburst/edit is an image to image tool from fal that runs in the background and returns images. Pay as you go from $0.0000065 per token, with one API key.

Try itfrom $0.0000065 / token

Pricing

Up to $0.23153 per run at the default inputs. quality: auto $0.92534, low $0.026, medium $0.05811, high $0.23153, xhigh $0.41145, max $0.92534. $39.00 per 1M image output tokens, $6.50 per 1M text input tokens, $10.40 per 1M image input tokens. The final charge comes from usage.

Limits

Runs: In the background
Returns: image
Longest queue wait: 3600s
Run deadline: 1800s
Largest inline input: 1024 KB
Largest input file: 500 MB
Largest output: 4768 MB
Outputs per run: 16

Input

Field	Kind	Default	Description
`prompt` required	text	—	The prompt for image generation
`image_urls` required	URLs or files	—	The URLs of the images to use as a reference for the generation. A maximum of 16 images are allowed.
`mask_url`	URL or file	—	Mask image URL. Standard processing treats white as edit and black as keep. With image_passthrough or provider_options, supply an OpenAI-native PNG alpha mask: transparent regions are edited, and dimensions must match the first reference.
`image_size`	JSON	auto	The size of the generated image. Use 'auto' to infer from input images.
`background`	one of auto, transparent, opaque	auto	Background for the generated image
`quality`	one of auto, low, medium, high, xhigh, max	high	Quality for the generated image. Higher settings increase detail, latency, and token usage. Use 'auto' to let the model choose.
`num_images`	number (whole) 1 to 10	1	Number of images to generate
`output_format`	one of jpeg, png, webp	png	Output format for the images
`output_compression`	number (whole) 0 to 100	—	Compression level from 0 to 100. Only supported when output_format is 'jpeg' or 'webp'.
`image_passthrough`	true or false	false	Send reference images and masks without resizing or re-encoding, and return the provider's original image bytes and format. Requests use the direct OpenAI API. Masks must be PNGs with an alpha channel: transparent regions are edited, and dimensions must match the first reference.

Same request in code

curl https://api.spiral.ai/v1/runs \
  -H "Authorization: Bearer $SPIRAL_API_KEY" \
  -H "Content-Type: application/json" \
  -H "Idempotency-Key: $(uuidgen)" \
  -d '{
    "provider": "fal",
    "model": "openai/gpt-image-2.5/sunburst/edit",
    "input": {
      "prompt": "Same workers, same beam, same lunch boxes - but they'\''re all on their phones now. One is taking a selfie. One is on a call looking annoyed. Same danger, new…",
      "image_urls": [
        "<image_urls>"
      ]
    }
  }'