The pricing standards for image generation models include two categories: Image tokens and Image generation. While the input pricing within Image tokens is straightforward, what confuses me is the difference between Image generation and output within Image tokens.
These are billed at the output image token price for the model, resulting in an output cost:
Per-image output pricing by size (n=1)
Model
Quality
1024×1024
1024×1536
1536×1024
GPT Image 1.5
Low
$0.008704
$0.013056
$0.012800
Medium
$0.033792
$0.050688
$0.050176
High
$0.133120
$0.199680
$0.198656
Model
Quality
1024×1024
1024×1536
1536×1024
GPT Image 1
Low
$0.01088
$0.01632
$0.01600
Medium
$0.04224
$0.06336
$0.06272
High
$0.16640
$0.24960
$0.24832
Model
Quality
1024×1024
1024×1536
1536×1024
GPT Image 1 Mini
Low
$0.002176
$0.003264
$0.003200
Medium
$0.008448
$0.012672
$0.012544
High
$0.033280
$0.049920
$0.049664
Output pricing per 1000 images by size (1k)
in dollars and cents for easier understanding
Model
Quality
1024×1024
1024×1536
1536×1024
GPT Image 1.5
Low
$8.71
$13.06
$12.80
Medium
$33.80
$50.69
$50.18
High
$133.12
$199.68
$198.66
Model
Quality
1024×1024
1024×1536
1536×1024
GPT Image 1
Low
$10.88
$16.32
$16.00
Medium
$42.24
$63.36
$62.72
High
$166.40
$249.60
$248.32
Model
Quality
1024×1024
1024×1536
1536×1024
GPT Image 1 Mini
Low
$2.18
$3.27
$3.20
Medium
$8.45
$12.68
$12.55
High
$33.28
$49.92
$49.67
Input
Note that image input (for the image AI to look at), in a chat with the image tool called or images as input on the edits endpoint, can cost you $0.035 to $0.067per image on gpt-image-1.5, or on gpt-image-1 when using “input_fidelity”: “high” - the big surprise pricing.