Positioning
Gemini Omni Flash is the first model in Google’s Gemini Omni family. It combines Gemini’s multimodal intelligence with generative media capabilities, initially focusing on video creation and conversational editing. (blog.google)
Capabilities and uses
- Accepts text, images, audio and video as inputs, and generates video with audio. (storage.googleapis.com)
- Supports natural-language editing across multiple turns, such as changing a scene’s environment, camera angle, style, characters or objects. It can be used for creative video production and iterative editing. (blog.google)
- Google says generated videos include an imperceptible SynthID watermark. (blog.google)
Limitations
Google’s model card notes that complete consistency across edits, complex motion and perfectly accurate rendered text remain challenges. At launch, video was the initial output focus; additional output modalities were described as future capabilities, not as launch features. (storage.googleapis.com)
Release and availability
Announced and initially released on May 19, 2026 at Google I/O. The initial rollout included the Gemini app and Google Flow for Google AI Plus, Pro and Ultra subscribers, with YouTube availability also announced. Google announced developer access through Google AI Studio, the Gemini API and Gemini Enterprise Agent Platform on June 30, 2026. (blog.google)
Pricing
Google Cloud Gemini Enterprise Agent Platform, as listed for Gemini Omni Flash and Gemini Omni 1.1 Flash; prices in USD per 1 million tokens, checked November 16, 2026:
- Input, including text, image, video and audio: $1.50.
- Text output, including response and reasoning: $9.00.
- Video output: $17.50.
Google lists input tokenization of 1,120 tokens per image, 32 tokens per second of audio, and up to 5,792 tokens per second of video. For video output with audio, it lists 1,931 tokens per second at 360p, 5,792 at 720p, 8,688 at 1080p and 17,376 at 4K. These are token-based charges, not a flat per-video fee. The cited table also lists separate upscaling token rules. Regional, cache and batch conditions for this model are not verified here. (cloud.google.com)