Google released Nano Banana 2 Lite, an image model that generates text-to-image outputs in roughly 4 seconds, and opened access to Gemini Omni Flash for video generation and conversational editing. Both are available now via the Gemini API and Google AI Studio, giving product teams faster and cheaper ways to prototype and scale multimedia features.
Nano Banana 2 Lite cuts image generation costs
Nano Banana 2 Lite is built for high throughput and quick iteration. It creates images at 1K resolution for $0.034 each, a price point that makes rapid visual drafting affordable. Despite the speed, the model holds prompt adherence and character consistency well.
The model is accessible in Google AI Studio, the Gemini API, and the Gemini Enterprise Agent Platform. It is also being integrated into consumer-facing surfaces like AI Mode in Search and the Gemini app.
Gemini Omni Flash handles video creation and editing
Gemini Omni Flash, first shown at Google I/O, now lets developers generate and edit video using text, image, or video inputs. It costs $0.10 per second of output video. The model refines videos through natural language instructions, uses multimodal references to maintain scene coherence, and draws on Gemini's real-world knowledge.
Currently, Omni Flash produces videos up to 10 seconds long. Audio references and scene extension are not yet supported in the API. Developers getting started with video generation can consult Generative Video Training for step-by-step guidance.
Chaining models for end-to-end workflows
Google designed these tools to work together. A product team can first generate concept art with Nano Banana 2 Lite, then feed those still images into Omni Flash to create animated short-form videos. Two demo applications illustrate this: "Anywhere" transports users to landmarks via generated images that Omni Flash animates, while "Space Lift" reimagines room interiors and turns them into cinematic video showcases.
Both models embed SynthID watermarking, which lets platforms verify AI-generated content. That built-in transparency helps teams manage content provenance without extra tooling.
Why this matters for product development
Speed and cost often dictate how many experiments a team can run. With image generation at $0.034 per output and video editing at $0.10 per second, product developers can iterate on creative concepts far more often without blowing the budget. The 4-second turnaround on still images shrinks feedback loops during design sprints, while natural-language video editing lets non-specialists produce polished motion content. Product teams can explore AI for Product Development Courses & Certifications to build the skills needed to embed these models directly into their build-measure-learn cycles.
Your membership also unlocks:
