Google Omni 1.1 Flash: 10 seconds at a time and up to 40 seconds at a time

Google launched Gemini Omni 1.1 Flash, a new version of Gemini Omni Flash, which was launched at the end of June with a full API coverage. A new version of the single-time video generation remains at a maximum of 10 seconds, but for the first time video continuation has been added to Omni API. Once a video is produced, it can now be taken back. The entire video can be taken for another 10 seconds at a time, with a maximum of 40 seconds after consecutive renewal. Each time you go back, the model takes into account a maximum of 10 seconds in front, so that the person, the action and the plot can be connected as much as possible. Omni 1.1 has also added front-end frame control. The user can give it an opening and an end picture, so that the model automatically produces a continuous video in the middle. For example, when people stand far away at the start of a given shot, when they end up in front of the camera, the model completes its own exercise. This time, a cheaper draft 360p model was added, which officially stated that the system had increased by up to 60 per cent, at a cost of about one third of 720p. After the draft, you can zoom in to 1080p or 4K. The 1080p and 4K here are magnifying outputs, not proto-high resolution generation。