Gemini Omni 1.2 Flash

Video generation and editing; successor to gemini-omni-1.1-flash, which Google auto-routes here from December 3, 2026. Google's changelog entry of November 5, 2026 announces it generally available, with up to 15-second video generation and conversational editing (3-10s for first and last frame interpolation, 3-10s per-turn video extension up to 30s total), better character and subject consistency across up to 6 reference images, 30% lower video output pricing and doubled paid-tier rate limits. The model page Versions block reads 'Stable: gemini-omni-1.2-flash'; the deprecations table gives release date November 5, 2026 and 'No shutdown date announced'. Paid tier only: the pricing page states 'Not available' on the free tier. Input $1.50 per 1M (labelled text / image / video). Two output rates (including thinking tokens): text $9.00/1M, video $12.00/1M — price_output_per_mtok holds the video rate, the same convention as gemini-omni-1.1-flash. Google bills on total output tokens at 5,792 tokens per second of 720p video, an effective ~$0.07/s (~$0.70 for a 10s clip); it also quotes $0.023/s at 360p (1,931 tokens/s), $0.104/s at 1080p with upsampling (8,688 tokens/s) and $0.209/s at 4K with upsampling (17,376 tokens/s). Limits per the model page: input 131,072 tokens, output 65,536 tokens, output video 3s-15s at 360p/720p/1080p/4K, 24 FPS. Model page lists inputs Text, Image, Video. No context caching rate is published. Read first-party 2026-11-07 from aistudio.google.com/docs (pricing, model page, deprecations, changelog).

Source: Google official documentation ↗

Explore more

Gemini Omni 1.2 Flash — questions & answers