Google has unveiled an updated AI model, Gemini Omni 1.2 Flash, focused on generating and editing video content. The new version expands video creation capabilities while simultaneously making the process significantly cheaper for developers and commercial users.
Key Improvements
The main technical achievement is an increase in the maximum length of generated video from 10 to 15 seconds. At the same time, the engineers focused not only on quantitative metrics but also on the quality of the output:
- consistency of characters and objects between frames has been improved, which is critical for narrative scenes;
- speech synchronization with the visuals has been enhanced;
- sequential editing via text prompts has been implemented, allowing targeted adjustments to be made without rebuilding the entire video.
Generation Economics
The pricing policy deserves particular attention. The cost of generation has been reduced by 30% — to $0.07 per second of video at 720p resolution. This move intensifies competition in the generative video market, where the price factor is becoming decisive for the mass adoption of the technology in production pipelines.
The Gemini Omni 1.2 Flash update reflects the broader trend toward cheaper and more controllable generative models. Lowering the cost per second of video while simultaneously increasing duration and quality opens the door to more complex use cases — from marketing content to prototyping video sequences in media production.
Cryptalist Analysis: A 30% reduction in generation price alongside expanded functionality indicates that the generative video market is entering a phase of price competition and infrastructure cost optimization. For institutional players, this is a signal of impending cost reductions in media and advertising production processes, which could boost venture capital interest in applied AI services.