Overview

• Veo 3.1 adds stronger prompt adherence, improved realism and richer background audio for more accurate, lifelike clips.
• The model outputs both horizontal 16:9 and vertical 9:16 video, expanding beyond Veo 3’s earlier landscape-only limitation.
• A new Veo 3.1 Fast tier targets cheaper, quicker generations and is available alongside the full model in the Gemini app, Gemini API and Vertex AI.
• Flow gains new controls including object insertion now, object removal coming soon, scene extension, multi-image guidance and first-to-last frame transitions with audio continuity.
• Coverage notes Google’s YouTube video trove has aided training progress, and prior statements pointed to bringing Veo tools to YouTube Shorts’ vertical format.