google/veo-3.1-fast-i2v

veo-3.1-fast-i2v
Docs

Google Veo 3.1 Fast image-to-video model. Supports two input modes: `image` mode (with optional `lastFrame` and `negativePrompt`) or `referenceImages`-only mode. Resolution up to 4k. See each parameter's description for detailed constraints.

$0.1200~$0.3500/sec
image-to-video

Input

Text describing what to avoid in the generated video. **Important:** Only supported when using `image` mode. NOT supported when using `referenceImages` mode — Google will return 400 error "Negative prompt is not supported in your use case".
Video description text guiding the animation
Video aspect ratio
16:9
Video duration in seconds (string type). **Constraint:** When using `lastFrame` or `referenceImages`, must be `8` — values `4` or `6` will be rejected by Google. When not using these features, `4`, `6`, `8` are all valid.
durationSeconds
Starting frame. Accepts an image URL (e.g. `https://example.com/frame.jpg`). Required unless using `referenceImages`-only mode. Cannot be combined with `referenceImages`.
Hint: Drag and drop files, paste from clipboard (Ctrl/Cmd+V), or provide a URL.
End frame. Accepts an image URL. **Constraints:** Requires `image` (cannot be used alone). Cannot be combined with `referenceImages`. `durationSeconds` must be `8` when using this parameter.
Hint: Drag and drop files, paste from clipboard (Ctrl/Cmd+V), or provide a URL.
Person generation policy
personGeneration
Reference images for style/character/visual guidance (up to 3). Each item is an image URL (e.g. `https://example.com/ref.jpg`). Supported only by Veo 3.1 and Veo 3.1 Fast I2V. Can be used without `image` (referenceImages-only mode). **Constraints:** Cannot be combined with `image` or `lastFrame`. Cannot be combined with `negativePrompt`. `durationSeconds` must be `8` when using this parameter.
Hint: Drag and drop files, paste from clipboard (Ctrl/Cmd+V), or provide a URL.
Video resolution. **Constraint:** 1080p and 4k are only available when `durationSeconds` is `8`.
resolution

Result

No results yet

Run the model to preview the output here.

More in this series