Google’s new anything-to-anything AI model is wild

A reviewer tested Google’s new Gemini Omni Flash model in Flow with an AI-generated deer character.

It improved on Veo in some edits and character consistency, but still produced odd visual glitches and inconsistent object behavior.

The model can handle text and video inputs for generation and editing, but results remain uneven.

Iterating is costly, making trial-and-error use expensive for creators.