Vercel Adds Gemini Omni 1.2 Flash Video Generation to AI Gateway
Vercel has added Google’s Gemini Omni 1.2 Flash to AI Gateway, allowing developers to generate and edit videos through one API key without a separate Google account. The model supports native audio, reference images, video extensions and resolutions up to 4K.
- Gemini Omni 1.2 Flash is now available on Vercel AI Gateway.
- It generates 3, 15 second videos with native audio.
- Videos can be edited with up to six reference images and extended up to 30 seconds total.
- Output supports 360p, 720p, 1080p and 4K in 16:9 or 9:16.
- AI Gateway has no markup or platform fee on inference, according to Vercel.
Video generation and editing
Gemini Omni 1.2 Flash can generate videos from text prompts ranging from 3 to 15 seconds, with native audio. It can also edit source clips between 3 and 15 seconds using up to six reference images.
Developers can extend videos by 3 to 10 seconds per call, up to 30 seconds total. Generated clips can also be refined in follow-up calls by using a previousInteractionId.
Output is available at 360p, 720p, 1080p and 4K resolutions. The model supports both 16:9 and 9:16 formats, giving developers options for different video layouts.
Access through the AI SDK
Vercel’s announcement shows developers how to call the model with the google/gemini-omni-1.2-flash model name in the AI SDK. The generateText function can return the generated video as a file, which developers can save as an MP4.
The example requests a three-second, 360p video based on the prompt, “A red ball rolls across a white table.” It asks for text and video response modalities, then finds the returned file by its video/mp4 media type.
For streaming workflows, developers can use streamText and read fullStream as results arrive. Vercel says the video is delivered as a completed file, and the example writes it to video.mp4 when the corresponding file part appears.
What AI Gateway provides
AI Gateway provides a unified API for calling models, tracking usage and cost, and configuring retries, failover and performance optimizations for higher-than-provider uptime, according to Vercel.
The service also includes custom reporting, Zero Data Retention support, budgets for API keys and routing rules. Vercel says AI Gateway reflects provider pricing with no markup and does not charge a platform fee on inference, including requests made through Bring Your Own Key, or BYOK.
For developers, the addition places Google’s video model alongside other models in the AI Gateway catalog while keeping access, usage tracking and configuration within the same gateway. Vercel directs users to the catalog to explore more models.
Vercel: Gemini Omni 1.2 Flash now available on AI Gateway
