Pay as you go

Pricing is per image generated. Charges are based on the resolution tier selected in the request, regardless of the resolution at which the image is processed internally, and apply equally to text-to-image and image-to-image generations.

Reference images, whether provided by the user or retrieved by the model, and prompts are included at no extra charge. The per-image rate does not vary with the number of reference images provided, up to the maximum set out in the documentation.

The resolution tier is selected in the request; all supported aspect ratios within a tier are charged the same. Delivered dimensions may differ slightly from the nominal size (dimensions are rounded to multiples of 16). Supported resolutions, aspect ratios and reference image limits are set out in the documentation.

Reference images up to 16 MP are accepted (larger ones are rejected); images above 4 MP are downscaled to 4 MP before processing. Rejected requests, failed generations and outputs withheld by moderation are not charged.

Tiers denote the approximate pixel count of the output: 768²≈0.6 MP, 1K≈1 MP (1024×1024-equivalent), 1.5K≈2 MP, 2K≈4 MP, 4K≈16 MP.

Text/Image-to-Video: any generation conditioned solely on text and/or one or more still images, including reference images and keyframes.

Video-to-Video: any generation conditioned in whole or in part on video input, including continuing, extending, editing or restyling video.

Prices are per second of video output (“Video Second”). Partial seconds are rounded up to the next whole second of the output.

Rates per Video Second are determined by (a) the generation modality, (b) the model variant selected (e.g., FLUX 3 Video, FLUX 3 Video Draft), and (c) the resolution band of the Output.

Resolution bands are determined by the total number of pixels per frame in an Output, where one megapixel (MP) equals 1,048,576 pixels (1024×1024). HD: above 0.5 and up to 1.0 MP per frame (e.g. 1280×704, 960×960, 1440×608). Full HD: above 1.0 and up to 2.0 MP per frame (e.g. 1920×1088, 1440×1440, 2176×928). QHD (2K): above 2.0 and up to 4.0 MP per frame (e.g. 2560×1440). UHD (4K): above 4.0 and up to 8.0 MP per frame (e.g. 3840×2176, 2880×2880).

Each Output falls within exactly one resolution band regardless of aspect ratio.

FLUX 3 Video Draft is only available in HD

FLUX 3 Video Draft Enhance, a full-quality render of a prior draft generation, is priced as a regular generation for the specific mode and the resolution band selected

Audio is included at no extra charge.

Clips up to 20 seconds.

Pricing is per megapixel (MP) of image processed: input pricing per MP for each reference image provided for image editing operations, plus output pricing per MP of output resolution. One megapixel equals 1024 × 1024 pixels (1,048,576 pixels).

1 MP = 1024×1024 px.

Resolution is rounded up to the next MP, separately for each reference image and the generated output.

Resolution measurements are rounded up to the next whole MP, separately for input and output.

Multiple reference images each count as 1 MP.

Reference images: a single reference image up to 4 MP is processed at its original resolution; above 4 MP it is downsampled to and charged as 4 MP. When multiple reference images are provided, each image exceeding 1 MP is downscaled to 1 MP, and each reference image is charged as 1 MP regardless of the output resolution requested.

Maximum output resolution is four (4) MP for all operations.

Images exceeding 4 MP are resized to 4 MP.

Resolution is rounded up to the next MP, separately for each reference image and the generated output.

1 MP = 1024×1024 px.

Multiple reference images each count as 1 MP.

Images exceeding 4 MP are resized to 4 MP.

Each FLUX.1 image is generated with up to 1,048,576 pixels per image. Height and width are between 256 and 1,440 pixels, each a multiple of 32.

Pricing is per megapixel (MP) of image processed: input pricing per MP for each reference image provided for image editing operations, plus output pricing per MP of output resolution. One megapixel equals 1024 × 1024 pixels (1,048,576 pixels).

1 MP = 1024×1024 px.

Resolution is rounded up to the next MP, separately for each reference image and the generated output.

Resolution measurements are rounded up to the next whole MP, separately for input and output.

Multiple reference images each count as 1 MP.

Reference images: a single reference image up to 4 MP is processed at its original resolution; above 4 MP it is downsampled to and charged as 4 MP. When multiple reference images are provided, each image exceeding 1 MP is downscaled to 1 MP, and each reference image is charged as 1 MP regardless of the output resolution requested.

Maximum output resolution is four (4) MP for all operations.

Images exceeding 4 MP are resized to 4 MP.