Google: Veo 3.1

Google: Veo 3.1 is a video generation model from Google. It is listed on api14 under the video category, and the provider release date is 2026-03-23. Google describes it as its state-of-the-art video generation model, built for maximum visual fidelity in final production cuts.

The model takes text and image inputs and produces video as its only output. It supports resolutions of 720p, 1080p and 4K, aspect ratios of 16:9 and 9:16, and clip durations of 4, 6 or 8 seconds. It can generate native synchronized audio, and it accepts frame images as first_frame and last_frame, so you can anchor the start and end of a clip. There is no context length figure listed for this model, so treat the prompt as a creative brief rather than a long document to reason over.

Those capabilities point to a fairly specific set of jobs. Text to video suits short ad spots, social clips and concept visualisation where you describe a scene and get a finished shot back. Image to video, combined with first_frame and last_frame, fits shot-to-shot continuity work, animating a still, or bridging two keyframes in an edit. The 9:16 ratio covers vertical formats for mobile feeds, while 16:9 and 4K suit widescreen delivery. Native audio generation means dialogue, ambience or effects can arrive with the picture instead of being added in a separate pass, which shortens the pipeline for draft cuts.

You call it through api14's OpenAI-compatible API using the model id "google/veo-3.1", so existing OpenAI-style client code works with a change of base URL and model name. The supported request parameters are max_tokens, response_format, seed, temperature and top_p. Seed is useful when you want reproducible generations across retries, and temperature and top_p give you control over how much variation you get between runs. Use response_format to declare how you want the result returned.

One thing to plan around is that this is a generation model, not a chat or reasoning model. It has no context length listed, it does not accept audio or video as input, and its output is video only, so tasks like transcription, video editing of existing footage or text analysis belong with a different model. If your work is text or code rather than moving images, another model on api14 will fit better.