The news of November 6thGoogleOfficial launch Gemini Omni 1.2 Flash Video Generation ModelAnd Gemini API and Gemini for business The platform is fully accessible。
A new version of the model was achievedOriginal 15 seconds video generation, multi-cycle interactive editing, enhanced reference compliance, and at the same time, more than 1.1 version30% reduction720p video generating price$ 0.07 per second(Note: The current exchange rate is approximately RMB 0.47。
InScale downOn the other hand, a new version of the workflow could be quickly drafted at 360p or 720p, with subsequent editing to amend the details without having to regenerate the entire segment and then upgrade the finished product to 1080p or 4K to further compress the cost of use。
InSnippet LengthOn top, Gemini Omni 1.2 Flash extends the length of the original-supported video from 10 seconds in version 1.1 to a maximum of 15 seconds, and supports the self-defined duration in 1 second, in 3 to 15 seconds, to complete the full story, product interpretation or social advertising clip generation once。
InReference ComplianceOn the other hand, users can combine role, product or style pictures, up to 3 seconds of dynamic clips and audio threads, and models can be used to keep face, clothing, packaging fonts, brand colours in line with different lenses throughout the journey, facilitating the conversion of existing product charts, role-setting maps to virtual trials, product rotations and brand video advertising。
InCoherence of resultsOn the other hand, the new version of the model, from the front to the end frame, allows for a strong alignment of roles, objects, scenes, and reduces visual drift and deformation during the motion and lens exchange; whether it is interpolated between the key frame at the end of the front or extended to build long sequences, the products, roles, props and the environment can be consistent between different lenses。
InDialogue EditorOn the other hand, users can adjust the same segment on a rotational basis, such as replacing products or contexts, modifying light or weather, changing image scales without having to be regenerated from scratch, and the multi-cycle dialogue editorial capacity can cover a complete 15-second session, facilitating the generation of an advertising variant for different audiences and places, without having to re-write and re-create。
InVoice EffectsOn the above, the new version of the model can generate a more fluid and natural dialogue, a clearer synchronization of the mouth, and support multi-talker condition control. It applies to multi-player dialogue in Flow, presentation video in Gemini for business and professional video in Google Vids with digitals, music and bystanders。
InFilm-level video controlUP THERE, USERS CAN DIRECT COMPLEX CONTINUOUS CAMERA MOVEMENTS, SUCH AS PANORAMA FPV CROSSING, DRAMATIZING AND REVOLVING, SIGNIFICANTLY REDUCING UNNECESSARY JUMP CUTS, AND MAINTAINING A MORE STABLE SCALE AND A MORE SOPHISTICATED FILM QUALITY。
Gemini Omni is now fully operational for the first time, with support from Google Cloud enterprise-level infrastructure and security capabilitiesPre-introduced ThroughputGenerate content that is invisible to the naked eye SynthID Watermark。
Developers can debug, view documents and pricing through the Google AI Studio prototype; business users can experience Google Cloud Conservator; ordinary users can direct scenes and multiple-wheel editors in Flow, generate free 15-second clips in Google Vids, Gemini App is sent gradually to Google AI Ultra subscriptions, Google Ads supports the placement of scaled video advertisements。