Google Rolls Out Nano Banana 2 Lite, Gemini Omni Flash

Google unveiled two new AI tools—Nano Banana 2 Lite and Gemini Omni Flash—on June 30, 2026, designed to streamline multimedia creation and editing for developers. These updates, part of Google DeepMind’s Gemini AI suite, focus on speed, cost-efficiency, and ease of use across image and video workflows.

Nano Banana 2 Lite, branded as Gemini 3.1 Flash Lite Image, is the faster, more cost-effective variant of the Nano Banana 2 model released in February 2026. This lightweight model targets developers seeking high-throughput, real-time image generation without sacrificing quality. With a latency of just 4 seconds per image and a cost of $0.034 per 1,000 images, it’s built for rapid prototyping and budget-conscious workflows. Key platforms such as Google AI Studio, Gemini API, and Gemini Enterprise Agent Platform already support the model, with integrations into consumer-facing tools like Search AI Mode and Google Photos rolling out simultaneously.

Performance benchmarks show Nano Banana 2 Lite excels in interactive use cases, offering prompt adherence, character consistency, and readable in-image text. Google has positioned it as the replacement for the original Nano Banana 2.5 Flash Image, highlighting its superior efficiency and scalability for developers requiring quick iterations or managing operational costs.

Meanwhile, Gemini Omni Flash, a multimodal AI tool for video generation and conversational editing, is now available to developers through the same platforms. This model enables users to create high-quality videos by combining inputs like text, images, and existing video clips. Priced at $0.10 per second of video output, it matches the cost of Veo 3.1 Fast, a competitor in the video AI space. Omni Flash’s standout features include conversational video editing using natural language commands, seamless synchronization of text and video actions, and the ability to draw on real-world knowledge for storytelling and scene construction.

Omni Flash currently supports 10-second video outputs, with longer durations expected in future updates. Although some features, such as audio references and extended scene continuity, remain in development, the model provides a compelling tool for developers looking to integrate AI-driven video workflows.

The real innovation lies in combining the two models. Developers can generate images with Nano Banana 2 Lite and use them as references in Gemini Omni Flash to produce animated clips, enabling end-to-end creative workflows. Google has showcased this synergy through demo apps like Anywhere, which transforms selfies into landmark animations, and Space Lift, which reimagines interiors with AI-driven design and cinematic video previews.

Both models integrate SynthID watermarking for content authentication, reinforcing Google’s commitment to safety and transparency in AI-generated media. Developers can access these tools starting today in Google AI Studio and via the Gemini API. With AI’s growing role in creative industries, these releases underline Google’s focus on scalable, accessible generative media solutions.

Image source: Shutterstock