Google DeepMind has made its Gemini Omni 1.2 Flash model generally available, introducing longer video generation capabilities and lower pricing. As reported by The Decoder, the updated model can now produce video clips up to 15 seconds in length. It also supports conversational, step-by-step editing and is priced at $0.07 per second for 720p output, representing a 30 percent price cut compared to the previous version. Users can also draft videos in lower resolutions and upscale only the final output to reduce costs further.
Expanding Capabilities in Video Generation
AI video generation models have advanced rapidly over the past year, moving from short, erratic seconds of footage to more coherent clips with longer runtimes and interactive editing features. By enabling step-by-step conversational edits, Google is targeting workflows where creators want to iteratively refine generated scenes rather than accept a single output. Lowering the cost threshold to $0.07 per second aligns with a broader industry trend toward cheaper inference, making high-resolution video synthesis more accessible for routine testing and prototyping.
Unanswered Questions and What Lies Ahead
Despite the general availability announcement, several operational details remain absent. The announcement did not specify exact hardware requirements or latency metrics for generating the 15-second clips. Furthermore, Google did not disclose specific details regarding the training data used for this version, nor did it outline rollout schedules for enterprise versus consumer tiers. Observers will need to monitor how real-world performance handles complex prompts and whether conversational editing holds up across multi-step iterations.
Further updates from Google DeepMind will clarify how these pricing structures and editing tools scale under heavy production workloads.
Lead image: Francesco Paggiaro / Pexels
