Google's Gemini Omni 1.2 Flash offers 15-second video creation, multi-turn editing, and enhanced reference adherence at 30% lower cost than its predecessor.

Google has announced the release of Gemini Omni 1.2 Flash, an update to its multimodal video-generation and conversational-editing platform, now available through the Gemini API and Google Cloud. The new version introduces native 15-second video creation, multi-turn conversational editing, and improved reference adherence, while reducing costs by 30% compared to its predecessor, Gemini Omni 1.1 Flash.

Gemini Omni 1.2 Flash enables users to generate and edit videos up to 15 seconds long, a 50% increase from the previous 10-second limit. Users can specify durations in one-second increments from 3 to 15 seconds, allowing for more flexibility in storytelling, product explainers, and social media advertisements. The pricing for 720p video output is set at $0.07 per second, making a 15-second clip cost $1.05 before additional charges.

The platform also enhances reference adherence, maintaining consistency in faces, garments, typography, and brand colors across camera angles. For example, Google demonstrated the generation of a 15-second virtual try-on sequence that preserved a model's identity and garment details across 360-degree turns and match cuts.

Multi-turn conversational editing allows users to refine clips iteratively without starting from scratch. This feature supports adjustments such as swapping backgrounds, changing lighting, or altering aspect ratios, making it easier to produce ad variants for different audiences and platforms. In one demonstration, Google showcased how a single clip could be modified turn by turn to feature different settings while maintaining character and scene consistency.

Source link