We use cookies. Cookies and browser storage keep you signed in and remember your preferences. With your permission we also measure how the site is used. You can any time. See our cookie policy.
google/gemini-omni-1.2-flashRuns on your balance at the prices above. New accounts start with free usage.
Settings
Gemini Omni 1.2 Flash is a native multimodal model from Google, positioned as the successor to Gemini Omni 1.1 Flash. It sits in the video category while also handling text, so it is built for work that combines language understanding with video generation and editing rather than for text alone.
The model accepts text, images and video as input, and it can return text or video as output. Its context window is 131,072 tokens, with a maximum output of 65,536 tokens. On the video side it supports resolutions of 360p, 720p, 1080p and 4K, aspect ratios of 16:9 and 9:16, and clip durations from 3 to 15 seconds. You can supply a first frame, a last frame, or both, and audio generation is available. Reasoning controls are part of the parameter set, which means the model can expose its reasoning process when you ask for it.
Those capabilities point to a fairly specific set of jobs. Because it takes video as input as well as producing it, it fits editing and transformation work: restyling existing footage, extending or adjusting a clip, or regenerating segments while keeping the surrounding material consistent. The first frame and last frame inputs make it useful for controlled transitions and for matching a shot to a planned start and end. The two aspect ratios cover both landscape delivery and vertical formats, and the resolution range lets you iterate quickly at lower resolutions before rendering a final version at 1080p or 4K. Audio generation means you can produce a complete short clip without a separate sound pass. Text input alongside video output also suits prompt driven storyboard or concept work, where you describe a scene and get a short clip back.
You call the model through api14's OpenAI-compatible API using the model id "google/gemini-omni-1.2-flash". The supported request parameters are include_reasoning, max_tokens, reasoning, reasoning_effort, seed, temperature and top_p. The seed parameter is worth using when you want reproducible output across runs, and temperature and top_p give you control over how much variation you get between generations. The reasoning parameters let you decide whether reasoning is surfaced and how much effort is spent on it.
Keep the output ceiling in mind: 65,536 tokens is the maximum output, and clips are capped at 15 seconds, so longer sequences need to be built from multiple generations. If your task is purely text generation with no video component, a text focused model is likely a simpler fit. For work that mixes language with short video generation or editing, this model covers the full path from prompt to finished clip.
import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://api14.com/api/v1",
apiKey: process.env.API14_API_KEY,
});
// Video generation is asynchronous: submit, then poll until completed.
let job = await client.post("/videos", { body: {
"model": "google/gemini-omni-1.2-flash",
"prompt": "A drone shot over turquoise ocean waves at sunrise",
"duration": 8,
"resolution": "720p",
"aspect_ratio": "16:9"
} });
while (!["completed", "failed"].includes(job.status)) {
await new Promise((r) => setTimeout(r, 5000));
job = await client.get(`/videos/${job.id}`);
}
console.log(job.status, job.unsigned_urls);
include_reasoningmax_tokensreasoningreasoning_effortseedtemperaturetop_pThis community has no topics yet. Open the first one about Google: Gemini Omni 1.2 Flash.