Google's Gemini Omni 1.1 Flash tops video generation Arena with new director-like controls
新智元 · wechat · 2026-08-29
Google updated Gemini Omni 1.1 Flash, ranking first in text-to-video and second in image-to-video on the Arena. The update shifts focus from quality to control, introducing:
- Scene Extension: Ability to review up to 10 seconds of context for seamless continuation up to 40 seconds.
- First & Last Frame Control: Generate intermediate motion between specified start and end frames.
- Video Reference: Upload up to 3 seconds of reference video to guide motion and rhythm.
- 360p Draft Mode: 60% faster and 1/3 the cost of 720p for rapid iteration before upscaling to 4K.
- Pricing: Ranges from $0.03/s (360p) to $0.30/s (4K).
Related event: Gemini Omni 1.1 Flash Tops Text-to-Video Arena Rankings(4 posts)→
More from Multimodal
- Ironclad law of AI: The impossible is inevitable — jachiam0 · 2026-08-29
- Frustrated with manual prompting for large scenes, desires reference features — cocktailpeanut · 2026-08-29
- AI Country Singer generated with MiniMax H3 Motion Control — SawyerCroft777 · 2026-08-29
- GNODE wraps your ComfyUI nodes into a clean mini UI with only chosen params exposed — neuroform · 2026-08-29
- 4 Claude agents competed to build a music video editor, judged blindly by 3 more — BuffaloConscious7919 · 2026-08-29
- Apple Releases Luce for Relightable 3D Asset Generation — apple · 2026-08-29