Gemini Omni 1.2 Flash: Fast, affordable and supports longer clips
gemini
| Source: Google DeepMind | Original article
Google's Gemini Omni 1.2 Flash delivers faster, cheaper AI generation with longer 15‑second clips and stronger consistency.
Google has rolled out the general‑availability version of Gemini Omni 1.2 Flash, its latest AI video‑generation model. The upgrade pushes the maximum clip length from ten to fifteen seconds, trims the price to $0.07 per second of 720p output – about 30 % cheaper than the preceding 1.1 Flash – and promises tighter adherence to visual references, keeping characters, products and brand details consistent across shots.
The new workflow is designed for efficiency. Users can draft a video at low resolution (360p or 720p), then make a targeted edit – for example correcting a single detail – without re‑rendering the entire sequence. The model also supports conversational editing, letting creators ask the system to modify a scene in natural language. Google ties the service to its Cloud platform, positioning it as a production‑scale tool for marketers, game developers and other content creators who need rapid, affordable video output.
The move matters because generative video has remained costly and technically demanding. By lowering per‑second pricing and extending clip length, Gemini Omni 1.2 Flash narrows the gap between experimental demos and practical, high‑volume use cases. Faster, cheaper iterations could reshape how advertising agencies, e‑learning platforms and social‑media teams prototype visual content, potentially accelerating the shift toward AI‑first production pipelines.
Looking ahead, Google signals ongoing work on “fast sports action, rapid script changes and complex physics,” hinting at future models that handle longer runtimes, higher resolutions and more demanding visual effects. Observers will watch how quickly third‑party tools integrate the API, whether competitors can match the cost‑performance balance, and how the broader media ecosystem adapts to conversational video editing as a standard workflow.
Sources
Back to AIPULSEN