Published 8 April 2026
Veo 3.1 Lite
Model Cards are intended to provide essential information on Gemini models, including known limitations, mitigation approaches, and safety performance. Model cards may be updated from time-to-time; for example, to include updated evaluations as the model is improved or revised.
Published: April 2026
Model Information
Description
Veo 3.1 Lite is an addition to the Veo 3 series of video generation systems capable of synthesizing high-quality, high-resolution video with audio from a text prompt or input image.
Inputs
Natural-language text strings, such as instructions for creating a synthetic video using a visual description, and images.
Outputs
High-quality, high-resolution video with audio.
Architecture
Veo 3.1 Lite is based on Veo 3. For more information about the model architecture for Veo 3.1 Lite, see the Veo 3 model card.
Model Data
Training Dataset
Veo 3.1 Lite is based on Veo 3. For more information about the training dataset for Veo 3.1 Lite, see the Veo 3 model card.
Training Data Processing
For more information about the training data processing for Veo 3.1 Lite, see the Veo 3 model card.
Implementation and Sustainability
Hardware
Veo 3.1 Lite is based on Veo 3. For more information about the hardware used to train for Veo 3.1 Lite and Google’s continued commitment to operate sustainably for Veo 3.1 Lite, see the Veo 3 model card.
Software
Veo 3.1 Lite is based on Veo 3. For more information about the software for Veo 3.1 Lite, see the Veo 3 model card.
Distribution
Veo 3.1 Lite is distributed in the following channels; respective documentation shared in line:
Our models are available to downstream providers via an application program interface (API) and subject to relevant terms of use. There is no required hardware or software to use the model. For AI Studio and Gemini API, see the Gemini API Additional Terms of Service. For more information, see Gemini Model API instructions.
Evaluation
Approach
We evaluated Veo 3.1 Lite vs Veo 3.1 Fast across both T2V (prompt-to-video generation) and I2V (image-to-video generation).
Results
Veo 3.1 Lite T2V achieved 54.6% overall win-rate across 1000 prompts and I2V achieved 47.2% overall win-rate across 646 prompts.
Veo 3.1 Lite is based on Veo 3. For more information about the evaluation approach for Veo 3.1 Lite, see the Veo 3 model card.
Intended Usage and Limitations
Benefit and Intended Usage
For more information about the benefit and intended usage for Veo 3.1 Lite, see the Veo 3 model card.
Known Limitations
For more information about the known limitations for Veo 3.1 Lite, see the Veo 3 model card.
Acceptable Usage
For more information about the acceptable usage for Veo 3.1 Lite, see the Veo 3 model card.
Ethics and Safety
For more information about the responsibility and safety evaluation approach for Veo 3.1 Lite, see the Veo 3 model card. Based on our safety evals, we did not see safety regression in Veo 3.1 Lite compared to Veo 3.1.