Video models

Choose an exact model name for Generate Video. Every request requires a non-empty prompt and a model. The tables below are generated from the app’s video catalog. They describe the settings accepted by the API’s model validation; provider availability and plan limits also apply.

Choose settings

Send duration as a string, such as "5" or "8s". A range means every whole second in that range. A comma-separated list means only those durations.
Send an explicitly listed resolution and aspectRatio when the model offers them. 2k and 4k use a lowercase k. Omit a setting marked “Provider-specific” or “From input”.
If you omit duration, the API selects five seconds when listed, otherwise the first configured duration, or five seconds when no duration list is configured. Provider-specific models may use their own output length. Explicit settings make requests more predictable.
“First image” requires image. A model that accepts a last image also accepts tailImage alongside image. Supply both when you want first-to-last-frame interpolation.
Reference models take referenceImages, or referenceVideos where listed. The reference tables below list optional audio inputs and limits.

Model suffixes are not a reliable way to select the input mode. Grok Imagine 1.5, Kling 2.5 Turbo Standard, and Seedance 1.0 Pro Fast require an image. Grok Imagine 1.5 Text and Seedance v1 Pro Fast take text.

Model catalog

Veo

Model	Input	Duration (seconds)	Resolution	Aspect ratio
`Veo 3.1 Fast`	Text	4, 6, 8	`720p`, `1080p`, `4k`	`16:9`, `9:16`
`Veo 3.1 Lite`	Text	4, 6, 8	`720p`, `1080p`	`16:9`, `9:16`
`Veo 3.1 Lite Image`	First image	8	`720p`, `1080p`	`16:9`, `9:16`
`Veo 3.1 Fast Image`	First image	8	`720p`, `1080p`, `4k`	`16:9`, `9:16`
`Veo 3.1`	Text	4, 6, 8	`720p`, `1080p`, `4k`	`16:9`, `9:16`
`Veo 3.1 Image`	First image	8	`720p`, `1080p`, `4k`	`16:9`, `9:16`
`Veo 3.1 First-Last Frame`	First image + optional last image	8	`720p`, `1080p`, `4k`	`16:9`, `9:16`
`Veo 3.1 Fast First-Last Frame`	First image + optional last image	8	`720p`, `1080p`, `4k`	`16:9`, `9:16`

Gemini Omni

Model	Input	Duration (seconds)	Resolution	Aspect ratio
`Gemini Omni Flash`	Text	4–10 (whole seconds)	Provider-specific	`16:9`, `9:16`
`Gemini Omni Flash Reference`	Reference images or video	Provider-specific	Provider-specific	From input
`Gemini Omni Flash 1.1`	Text	3–10 (whole seconds)	`360p`, `720p`, `1080p`, `4k`	`16:9`, `9:16`
`Gemini Omni Flash 1.1 Reference`	Reference images or video	3–10 (whole seconds)	`360p`, `720p`, `1080p`, `4k`	From input
`Gemini Omni Flash Image`	First image	4–10 (whole seconds)	Provider-specific	From input
`Gemini Omni Flash 1.1 Image`	First image	3–10 (whole seconds)	`360p`, `720p`, `1080p`, `4k`	From input
`Gemini Omni Flash 1.1 First-Last Frame`	First image + optional last image	3–10 (whole seconds)	`360p`, `720p`, `1080p`, `4k`	From input

MiniMax

Model	Input	Duration (seconds)	Resolution	Aspect ratio
`MiniMax H3 Max`	Text	5–15 (whole seconds)	`768p`, `480p`	`16:9`, `9:16`, `1:1`, `4:3`, `3:4`, `21:9`
`MiniMax H3 Max Reference`	Reference images or video	5–15 (whole seconds)	`768p`, `480p`	`16:9`, `9:16`, `1:1`, `4:3`, `3:4`, `21:9`
`MiniMax H3`	Text	4–15 (whole seconds)	`768p`, `2k`	`16:9`, `9:16`, `1:1`, `4:3`, `3:4`, `21:9`
`MiniMax H3 Reference`	Reference images or video	4–15 (whole seconds)	`768p`, `2k`	`16:9`, `9:16`, `1:1`, `4:3`, `3:4`, `21:9`
`MiniMax H3 Max Image`	First image + optional last image	5–15 (whole seconds)	`768p`, `480p`	From input
`MiniMax H3 Image`	First image + optional last image	4–15 (whole seconds)	`768p`, `2k`	`16:9`

Grok

Model	Input	Duration (seconds)	Resolution	Aspect ratio
`Grok Imagine`	Text	5–15 (whole seconds)	`480p`, `720p`	`16:9`, `4:3`, `1:1`, `9:16`, `3:4`, `3:2`, `2:3`
`Grok Imagine Reference`	Reference images	5–10 (whole seconds)	`480p`, `720p`	`16:9`, `4:3`, `1:1`, `9:16`, `3:4`, `3:2`, `2:3`
`Grok Imagine 1.5 Text`	Text	5–15 (whole seconds)	`480p`, `720p`, `1080p`	`16:9`, `4:3`, `1:1`, `9:16`, `3:4`, `3:2`, `2:3`
`Grok Imagine 1.5 Reference`	Reference images	5–10 (whole seconds)	`480p`, `720p`	`16:9`, `4:3`, `1:1`, `9:16`, `3:4`, `3:2`, `2:3`
`Grok Imagine Image`	First image	5–15 (whole seconds)	`480p`, `720p`	From input
`Grok Imagine 1.5`	First image	5–15 (whole seconds)	`480p`, `720p`, `1080p`	From input

Seedance

Model	Input	Duration (seconds)	Resolution	Aspect ratio
`Seedance 2.0 Fast`	Text	4–15 (whole seconds)	`480p`, `720p`, `1080p`, `4k`	`16:9`, `9:16`, `1:1`, `4:3`, `3:4`, `21:9`
`Seedance 2.5`	Text	4, 5, 6, 8, 10, 12, 15, 20, 25, 30	`480p`, `720p`	`16:9`, `9:16`, `1:1`, `4:3`, `3:4`, `21:9`
`Seedance 2.5 Turbo`	Text	4, 5, 6, 8, 10, 12, 15, 20, 25, 30	`720p`, `1080p`	`16:9`, `9:16`, `1:1`, `4:3`, `3:4`, `21:9`
`Seedance 2.5 Turbo Reference`	Reference images or video	4, 5, 6, 8, 10, 12, 15, 20, 25, 30	`720p`, `1080p`	`16:9`, `9:16`, `1:1`, `4:3`, `3:4`, `21:9`
`Seedance 2.5 Reference`	Reference images or video	4, 5, 6, 8, 10, 12, 15, 20, 25, 30	`480p`, `720p`	`16:9`, `9:16`, `1:1`, `4:3`, `3:4`, `21:9`
`Seedance 2.0 Turbo`	Text	4–15 (whole seconds)	`720p`, `1080p`	`16:9`, `9:16`, `1:1`, `4:3`, `3:4`, `21:9`
`Seedance 2.0`	Text	4–15 (whole seconds)	`480p`, `720p`, `1080p`, `4k`	`16:9`, `9:16`, `1:1`, `4:3`, `3:4`, `21:9`
`Seedance 2.0 Mini`	Text	4–15 (whole seconds)	`480p`, `720p`, `1080p`, `4k`	`16:9`, `9:16`, `1:1`, `4:3`, `3:4`, `21:9`
`Seedance 2.0 Reference`	Reference images or video	4–15 (whole seconds)	`480p`, `720p`, `1080p`, `4k`	`16:9`, `9:16`, `1:1`, `4:3`, `3:4`, `21:9`
`Seedance 2.0 Fast Reference`	Reference images or video	4–15 (whole seconds)	`480p`, `720p`, `1080p`, `4k`	`16:9`, `9:16`, `1:1`, `4:3`, `3:4`, `21:9`
`Seedance 2.0 Mini Reference`	Reference images or video	4–15 (whole seconds)	`480p`, `720p`, `1080p`, `4k`	`16:9`, `9:16`, `1:1`, `4:3`, `3:4`, `21:9`
`Seedance 2.0 Image`	First image + optional last image	4–15 (whole seconds)	`480p`, `720p`, `1080p`, `4k`	`16:9`, `9:16`, `1:1`, `4:3`, `3:4`, `21:9`
`Seedance 2.0 Fast Image`	First image + optional last image	4–15 (whole seconds)	`480p`, `720p`, `1080p`, `4k`	`16:9`, `9:16`, `1:1`, `4:3`, `3:4`, `21:9`
`Seedance 2.5 Turbo Image`	First image + optional last image	4, 5, 6, 8, 10, 12, 15, 20, 25, 30	`720p`, `1080p`	`16:9`
`Seedance 2.5 Image`	First image + optional last image	4, 5, 6, 8, 10, 12, 15, 20, 25, 30	`480p`, `720p`	`16:9`
`Seedance 2.0 Turbo Image`	First image + optional last image	4–15 (whole seconds)	`720p`, `1080p`	`16:9`, `9:16`, `1:1`, `4:3`, `3:4`, `21:9`
`Seedance 2.0 Mini Image`	First image + optional last image	4–15 (whole seconds)	`480p`, `720p`, `1080p`, `4k`	`16:9`, `9:16`, `1:1`, `4:3`, `3:4`, `21:9`
`Seedance 1.5 Pro`	Text	4–12 (whole seconds)	`480p`, `720p`, `1080p`	`16:9`, `4:3`, `1:1`, `3:4`, `9:16`, `21:9`
`Seedance 1.5 Pro Image`	First image	4–12 (whole seconds)	`480p`, `720p`, `1080p`	`16:9`, `4:3`, `1:1`, `3:4`, `9:16`, `21:9`
`Seedance 1.5 Pro First-Last Frame`	First image + optional last image	4–12 (whole seconds)	`480p`, `720p`, `1080p`	`16:9`, `4:3`, `1:1`, `3:4`, `9:16`, `21:9`
`Seedance v1 Pro Fast`	Text	3–12 (whole seconds)	Provider-specific	`1:1`, `3:4`, `9:16`, `16:9`, `4:3`, `21:9`
`Seedance 1.0 Pro Fast`	First image	3–12 (whole seconds)	Provider-specific	`1:1`, `3:4`, `9:16`, `16:9`, `4:3`, `21:9`

FLUX

Model	Input	Duration (seconds)	Resolution	Aspect ratio
`FLUX 3`	Text	5, 6, 8, 10, 12, 15, 20	`720p`, `1080p`	`16:9`, `9:16`, `1:1`, `4:3`, `3:4`, `21:9`
`FLUX 3 Image`	First image + optional last image	5, 6, 8, 10, 12, 15, 20	`720p`, `1080p`	`16:9`, `9:16`, `1:1`, `4:3`, `3:4`, `21:9`

WAN

Model	Input	Duration (seconds)	Resolution	Aspect ratio
`WAN 3.0`	Text	5, 8, 10, 12, 15, 20, 25, 30	`480p`, `720p`, `1080p`	`16:9`, `9:16`, `1:1`, `4:3`, `3:4`
`WAN 3.0 Reference`	Reference images	5, 8, 10, 12, 15, 20, 25, 30	`480p`, `720p`, `1080p`	`16:9`, `9:16`, `1:1`, `4:3`, `3:4`
`WAN 3.0 Image`	First image + optional last image	5, 8, 10, 12, 15, 20, 25, 30	`480p`, `720p`, `1080p`	`16:9`, `9:16`, `1:1`, `4:3`, `3:4`
`Wan 2.5`	Text	Provider-specific	Provider-specific	`16:9`, `9:16`, `1:1`
`Wan v2.2 Turbo`	Text	Provider-specific	Provider-specific	`16:9`, `9:16`
`Wan v2.2`	Text	Provider-specific	Provider-specific	`16:9`, `9:16`
`Wan 2.5 Image`	First image	Provider-specific	Provider-specific	`16:9`, `9:16`, `1:1`
`Wan v2.2 Turbo Image`	First image + optional last image	Provider-specific	Provider-specific	`16:9`, `9:16`
`Wan-2.2 Image`	First image + optional last image	Provider-specific	Provider-specific	`1:1`, `16:9`, `9:16`

Kling

Model	Input	Duration (seconds)	Resolution	Aspect ratio
`Kling 2.6 Pro`	Text	5, 10	Provider-specific	`16:9`, `9:16`, `1:1`
`Kling 2.6 Pro Image`	First image + optional last image	5, 10	Provider-specific	`16:9`, `9:16`, `1:1`
`Kling 3.0 Pro`	Text	3–15 (whole seconds)	Provider-specific	`16:9`, `9:16`, `1:1`
`Kling 3.0 Pro Image`	First image + optional last image	3–15 (whole seconds)	Provider-specific	`16:9`, `9:16`, `1:1`
`Kling 3.0 Standard`	Text	3–15 (whole seconds)	Provider-specific	`16:9`, `9:16`, `1:1`
`Kling 3.0 Standard Image`	First image + optional last image	3–15 (whole seconds)	Provider-specific	`16:9`, `9:16`, `1:1`
`Kling O3 Pro`	Text	3–15 (whole seconds)	Provider-specific	`16:9`, `9:16`, `1:1`
`Kling O3 Pro Image`	First image + optional last image	3–15 (whole seconds)	Provider-specific	`16:9`, `9:16`, `1:1`
`Kling O3 Standard`	Text	3–15 (whole seconds)	Provider-specific	`16:9`, `9:16`, `1:1`
`Kling O3 Standard Image`	First image + optional last image	3–15 (whole seconds)	Provider-specific	`16:9`, `9:16`, `1:1`
`Kling O3 Reference`	Reference images	3–15 (whole seconds)	Provider-specific	`16:9`, `9:16`, `1:1`
`Kling O1 Image`	First image + optional last image	5, 10	Provider-specific	`1:1`
`Kling O1 Reference`	Reference images	3, 5, 10	Provider-specific	`16:9`, `9:16`, `1:1`
`Kling 2.5 Turbo Pro Image`	First image + optional last image	5, 10	Provider-specific	`16:9`
`Kling 2.5 Turbo Pro`	Text	5, 10	Provider-specific	`16:9`, `9:16`, `1:1`
`Kling 2.5 Turbo Standard`	First image	5, 10	Provider-specific	`16:9`, `9:16`, `1:1`

Happy Horse

Model	Input	Duration (seconds)	Resolution	Aspect ratio
`Happy Horse`	Text	3–15 (whole seconds)	`720p`, `1080p`	`16:9`, `9:16`, `1:1`, `4:3`, `3:4`
`Happy Horse Image`	First image	3–15 (whole seconds)	`720p`, `1080p`	From input
`Happy Horse Reference`	Reference images	3–15 (whole seconds)	`720p`, `1080p`	`16:9`, `9:16`, `1:1`, `4:3`, `3:4`

LTX-2

Model	Input	Duration (seconds)	Resolution	Aspect ratio
`LTX-2 Pro`	Text	6, 8, 10	Provider-specific	`16:9`
`LTX-2 Fast`	Text	6, 8, 10, 12, 14, 16, 18, 20	Provider-specific	`16:9`
`LTX-2 Pro Image`	First image	6, 8, 10	Provider-specific	`16:9`
`LTX-2 Fast Image`	First image	6, 8, 10, 12, 14, 16, 18, 20	Provider-specific	`16:9`

Hailuo

Model	Input	Duration (seconds)	Resolution	Aspect ratio
`Hailuo 2.0`	Text	6, 10	Provider-specific	`16:9`
`Hailuo 2.0 Pro`	Text	Provider-specific	Provider-specific	`16:9`
`Hailuo 2.3`	Text	6, 10	Provider-specific	`16:9`
`Hailuo 2.3 Pro`	Text	5	Provider-specific	`16:9`
`Hailuo 2.3 Image`	First image	6, 10	Provider-specific	`16:9`
`Hailuo 2.3 Pro Image`	First image	5	Provider-specific	`16:9`
`Hailuo 2.3-Fast Pro Image`	First image	6	Provider-specific	`16:9`
`Hailuo 2.0 Image`	First image + optional last image	6, 10	Provider-specific	`16:9`
`Hailuo 2.0 Pro Image`	First image + optional last image	Provider-specific	Provider-specific	`16:9`

Pixverse

Model	Input	Duration (seconds)	Resolution	Aspect ratio
`Pixverse v5.5`	Text	5, 10	Provider-specific	`16:9`, `4:3`, `1:1`, `3:4`, `9:16`
`Pixverse v5.5 Image`	First image	5, 10	Provider-specific	`16:9`, `4:3`, `1:1`, `3:4`, `9:16`

Reference limits

All media URLs must be downloadable without browser cookies and remain valid through any queue delay. Limits are per request, not per account. Additional reference media can increase the credit cost. Do not combine referenceVideos with image or tailImage. Audio references require a companion reference image or a supported reference video; they cannot be used with a first-frame image.

Model	Reference images (max)	Reference video	Reference audio
`Veo 3.1 Fast`	3	—	—
`Gemini Omni Flash Reference`	5	1 max; 4–20s each; 20s total; 20 MiB each	—
`Gemini Omni Flash 1.1 Reference`	5	1 max; 4–10s each; 10s total; 20 MiB each	—
`MiniMax H3 Max Reference`	9	3 max; 2–15s each; 15s total; 20 MiB each	3 max; 2–15s each; 15s total
`Grok Imagine Reference`	7	—	—
`Grok Imagine 1.5 Reference`	7	—	—
`Seedance 2.5 Turbo Reference`	30	3 max; 2–30s each; 30s total; 50 MiB each	10 max; 2–30s each; 30s total; 15 MiB each; wav/mp3
`Seedance 2.5 Reference`	30	3 max; 2–30s each; 30s total; 50 MiB each	10 max; 2–30s each; 30s total; 15 MiB each; wav/mp3
`WAN 3.0 Reference`	9	—	3 max; 2–30s each; 30s total; 15 MiB each; wav/mp3
`Seedance 2.0 Reference`	9	3 max; 2–15s each; 15s total; 30 MiB each	3 max; 15s total
`Seedance 2.0 Fast Reference`	9	3 max; 2–15s each; 15s total; 30 MiB each	3 max; 15s total
`MiniMax H3 Reference`	9	3 max; 2–15s each; 15s total; 20 MiB each	3 max; 2–15s each; 15s total; 15 MiB each; wav/mp3
`Seedance 2.0 Mini Reference`	9	3 max; 2–15s each; 15s total; 30 MiB each	3 max; 15s total
`Veo 3.1 Fast Image`	3	—	—
`Happy Horse Reference`	9	—	—
`Kling O3 Reference`	4	—	—
`Kling O1 Reference`	4	—	—
`Veo 3.1`	3	—	—
`Veo 3.1 Image`	3	—	—
`Veo 3.1 First-Last Frame`	3	—	—
`Veo 3.1 Fast First-Last Frame`	3	—	—

Veo reference images are optional on the variants listed above. For Gemini Omni Flash Reference video editing, the output follows the input clip’s duration. Input requirements still apply when a model offers reference images alongside a first-frame mode.

Free plan settings

Free accounts can use the variants below with sufficient remaining credits. Send the listed resolution explicitly. All other models require a paid plan.

Model	Resolution	Maximum duration
`Veo 3.1 Fast`	`720p`	Model settings
`Seedance 2.0 Mini`	`480p`	5 seconds
`Seedance 2.0 Mini Reference`	`480p`	5 seconds
`Veo 3.1 Lite`	`720p`	Model settings
`Veo 3.1 Lite Image`	`720p`	Model settings
`Veo 3.1 Fast Image`	`720p`	Model settings
`Seedance 2.0 Mini Image`	`480p`	5 seconds
`Veo 3.1 Fast First-Last Frame`	`720p`	Model settings

Retired and separate features

Sora 2, Sora 2 Pro, and their Image variants are no longer accepted by this endpoint. Choose a current model and update its duration, resolution, and inputs together. Kling Motion Control uses a separate web-app workflow. Its motion-video payload is not supported by this public API endpoint.

Getting Started

Text Generation

Image Generation

Video Generation

Transcription

Text-to-Speech

Sound Effects

Chat Completions

Bots

Choose settings

Model catalog

Veo

Gemini Omni

MiniMax

Grok

Seedance

FLUX

WAN

Kling

Happy Horse

LTX-2

Hailuo

Pixverse

Reference limits

Free plan settings

Retired and separate features

​Choose settings

​Model catalog

​Veo

​Gemini Omni

​MiniMax

​Grok

​Seedance

​FLUX

​WAN

​Kling

​Happy Horse

​LTX-2

​Hailuo

​Pixverse

​Reference limits

​Free plan settings

​Retired and separate features

Choose settings

Model catalog

Veo

Gemini Omni

MiniMax

Grok

Seedance

FLUX

WAN

Kling

Happy Horse

LTX-2

Hailuo

Pixverse

Reference limits

Free plan settings

Retired and separate features