Google just changed what "AI video generation" means. Gemini Omni Flash doesn't generate a silent clip and hope you dub it later — it generates video and audio together, in one pass, synchronized from the start. Dialogue, ambient sound, Foley, music — all native, all grounded in Gemini's real-world knowledge instead of a text-to-video model guessing at physics.
Here's what Gemini Omni Flash actually does, what it costs through Google directly, and why creators are better off accessing it through a bundled studio than paying per-token.
Most AI video models — including earlier Veo versions — generate silent footage first. If you want sound, you add it afterward: a separate audio model, manual sync work, or a generic music bed slapped over the top. That's the workflow everyone's used to, and it's why AI-generated video so often feels slightly "off" — the audio and visual layers were never actually connected.
Gemini Omni Flash breaks that pattern. It's Google's first model in the new "Omni" family, launched to general availability on June 30, 2026, and it can take any combination of text, image, audio, and video as input and generate a new video where the sound and visuals are created together, in the same generation pass. Google describes it as a model that "can create anything from any input — starting with video."
Practically, that means:
It's also rolling out for free inside YouTube Shorts and the YouTube Create app, which tells you where Google thinks the mainstream use case is: fast, sound-native short-form content.
If you want Gemini Omni Flash through Google's own API, pricing is token-based, and it adds up faster than the headline number suggests:
| Item | Price |
|---|---|
| Input (text/image/video/audio) | $1.50 per 1M tokens |
| Text/thinking output | $9.00 per 1M tokens |
| Video output | $17.50 per 1M tokens |
| Effective 720p video cost | ~$0.10 per second of output |
| Effective 4K video cost | Significantly higher — 4K bills at ~9x the token rate of 720p |
At $0.10/second, a 10-second 720p clip runs roughly $1.00 in output tokens alone, before input tokens and any failed/discarded generations. Scale that across a real content calendar — a creator generating 20–30 clips a week to find the ones worth publishing — and direct API access gets expensive fast, on top of needing developer setup to use it at all (there's no simple consumer subscription; it's Google AI Studio or API-only).
This is the same pattern we've flagged with every frontier Google model: extraordinary capability, gated behind pricing and infrastructure built for developers, not creators paying out of pocket.
Gemini Omni Flash isn't a replacement for Veo 3.1 — it's a different tool for a different job. Veo 3.1 is Google's flagship model for cinematic quality and longer-form realism. Omni Flash is built for speed, native sound, and conversational back-and-forth editing on shorter clips.
| Model | Best for | Native audio | Access complexity |
|---|---|---|---|
| Gemini Omni Flash | Fast, sound-native short clips with iterative editing | Yes, from launch | API/dev-only via Google directly |
| Veo 3.1 | Cinematic quality, longer scenes | Partial (ambient/lip-sync) | API or Google AI Studio, $250+/mo territory |
| Seedance 2.0 / 2.5 | Multimodal reference-driven generation | No (visual only) | Varies by access point |
| Happy Horse (open-source) | Native joint audio-video, multilingual lip-sync | Yes, single-pass | Self-hosted/open weights |
The honest takeaway: native audio-video generation is becoming the norm across 2026's frontier models, not a Gemini-only feature. What still separates them is access — and that's the part creators actually feel in their budget.
Coverr added Gemini Omni Flash to its AI Video Generator the same week Google made it generally available — it's flagged live in the "What's New" rail on coverr.co/studio right now. Instead of setting up Google API billing, tracking token consumption across input/output/video categories, and paying per second, you generate inside Coverr's credit system alongside Veo 3.1, Kling 3.0, Sora, and Seedance 2.0 — from $4.20/mo, with 1,000 free renewable AI credits every month before you spend anything.
That matters most for the workflow Omni Flash is actually built for: fast iteration. If a model is designed for conversational, back-and-forth editing, you don't want per-token anxiety on every follow-up prompt. A flat monthly credit pool removes that friction entirely.
There's also Coverr's Recreate workflow: instead of prompting Omni Flash from a blank page, you can start from a real HD/4K clip in Coverr's stock library and use it as the reference input for a native audio-video generation — giving the model a real visual and tonal anchor instead of a guess.
Test Gemini Omni Flash's native audio generation against Veo 3.1 and Seedance 2.0 using Coverr's free monthly credits before deciding which model earns a spot in your regular workflow. For the full technical rundown on how Google prices Omni Flash at the API level, Google's official Gemini API pricing page has the current token rates.
Gemini Omni Flash is Google's first model in its new "Omni" family, generally available since June 30, 2026. It generates video and native, synchronized audio together in a single pass from text, image, audio, or video inputs, grounded in Gemini's world knowledge.
Google's Gemini API prices it at $1.50 per million input tokens, $17.50 per million video output tokens, and $9.00 per million text/thinking output tokens — working out to roughly $0.10 per second of 720p video output, before input costs.
Yes — native audio is a core feature, not an add-on. It generates dialogue, ambient sound, music, and timed sound effects synchronized with the video in the same generation pass.
They serve different purposes. Veo 3.1 targets cinematic quality and longer, more polished scenes. Gemini Omni Flash is optimized for speed, native audio, and fast conversational editing on shorter clips. Neither fully replaces the other.
Yes — Coverr added Gemini Omni Flash to its AI Video Generator, letting you generate with it through a monthly credit subscription starting at $4.20/mo instead of setting up Google API billing and per-token tracking.
Ready to try native audio-video generation without the token math? Generate your first Gemini Omni Flash clip free with Coverr's monthly AI credits.