Gemini Omni 1.2 Flash: Fast and cost-effective with longer clips
Gemini Omni 1.2 Flash adds native 15‑second video generation (up from 10 s) with one‑second granularity (3‑15 s), multi‑turn conversational editing, and stronger reference adherence, while cutting price by 30% to $0.07 per second for 720p. The model keeps character, object, and scene details consistent across frames, enabling efficient draft‑then‑edit workflows that can be upscaled to 1080p or 4K. Engineers can now produce longer, higher‑quality clips at lower cost through the Gemini API, Google Vids, and Gemini for Business. This reduces compute waste and simplifies production pipelines for video‑centric AI applications.
⚡ Key Takeaways
- Price reduced 30% to $0.07 per second for 720p video.
- Supports native video lengths up to 15 seconds, selectable in 1‑second increments from 3 to 15 seconds.
- Multi‑turn conversational editing lets users fix details without regenerating the whole clip.
- Consistency improvements maintain faces, garments, branding, and lighting across camera angles and frames.
- Workflow: draft at 360p/720p, edit via conversation, then upscale to 1080p or 4K.
🔧 Tools & Libraries
Engineers can deliver longer, brand‑consistent video ads and product demos at a fraction of prior cost, accelerating time‑to‑market for AI‑generated media pipelines.
✅ Practical Steps
- Call the Gemini API (or use Google Vids / Gemini for Business) to generate a video, specifying `duration` between 3 and 15 seconds.
- Use a follow‑up API request in the same conversation to edit the clip, adjusting details without re‑generating the full video.
- After finalizing, request an upscale to 1080p or 4K via the same API endpoint.
Want the full story? Read the original article.
Read on Google DeepMind ↗