Google Launches Gemini Omni Flash and Nano Banana 2 Lite to Deliver Faster, More Affordable AI Video and Image Creation

Introduction

The competition in generative AI is increasingly focused on one key challenge: delivering high-quality creative outputs without sacrificing speed or cost efficiency.

To address this demand, Google Cloud has announced the availability of two new models within the Gemini Enterprise Agent Platform: Nano Banana 2 Lite and Gemini Omni Flash. Together, these models aim to help businesses, developers, and creative professionals generate and edit images and videos faster while maintaining quality, consistency, and enterprise-grade governance.

The launch highlights Google's broader strategy of making advanced multimodal AI more practical for real-world production environments, from marketing and content creation to design workflows and AI-powered applications.

Gemini Omni Flash Brings Conversational Video Generation to Enterprises

One of the most significant announcements is the public preview of Gemini Omni Flash, a multimodal video generation and editing model built on Gemini's knowledge and reasoning capabilities.

Unlike traditional video editing tools that require complex timelines and manual adjustments, Gemini Omni Flash enables users to edit videos using natural language instructions.

Users can perform tasks such as:

• Swapping characters or products

• Relighting scenes

• Changing camera angles

• Applying style transfers

• Adding or removing objects

• Refining video content through conversation

The model also supports text, image, and video inputs, making it easier to create rich multimedia experiences inside agentic workflows and enterprise applications.

Four Key Capabilities Behind Gemini Omni Flash

Google highlighted four core strengths that differentiate Gemini Omni Flash from many existing video generation platforms.

Conversational Editing

Users can modify video content using simple natural language instructions while preserving existing audio and video tracks.

Multimodal Understanding

The model combines text, images, and videos as input references to guide generation and editing tasks.

Real-World Knowledge and Simulation

Leveraging Gemini's understanding of science, history, culture, and physics, the model can generate more contextually accurate and realistic content.

Text and Motion Synchronisation

Gemini Omni Flash can render readable text directly into video scenes while synchronising motion graphics and typography with on-screen actions.

This capability is particularly valuable for marketing videos, explainers, educational content, and branded media production.

Competitive Pricing for Video Generation

Cost remains one of the biggest barriers to large-scale AI video adoption.

Google says Gemini Omni Flash is priced at $0.10 per second of generated video output, positioning it among the strongest price-performance offerings in the current AI video market.

The company also announced that additional features will arrive soon, including:

• Audio references

• Video references

• Last-frame editing

• Scene extension capabilities

• Higher-resolution outputs

These enhancements are expected to further strengthen the platform's appeal for professional content creators and enterprise teams.

Industry Leaders Already Testing Gemini Omni Flash

Several major creative and marketing organisations have already integrated or tested Gemini Omni Flash.

Adobe

Adobe plans to bring Gemini Omni Flash and Nano Banana 2 Lite into Adobe Firefly, expanding creator flexibility and accelerating content production workflows.

Invideo

Invideo highlighted the model's visual effects capabilities and its potential to merge traditional production processes with AI-powered content creation.

WPP

Marketing giant WPP integrated Gemini Omni Flash into WPP Open, its agentic marketing platform, enabling asset localisation, product swaps, and dynamic style transfers.

These early partnerships demonstrate growing industry confidence in conversational AI-powered media production.

Nano Banana 2 Lite Focuses on Speed and Cost Efficiency

Alongside Gemini Omni Flash, Google announced the general availability of Nano Banana 2 Lite (Gemini 3.1 Flash-Lite Image).

The model is positioned as the fastest and most cost-efficient member of the Nano Banana image generation family.

According to Google, Nano Banana 2 Lite can generate images in as little as four seconds, making it ideal for:

• Rapid ideation

• Marketing asset creation

• Social media applications

• A/B testing campaigns

• Ecommerce experiences

• Real-time AI-powered design tools

The focus is clear: reduce waiting time and allow creators to iterate at the speed of their ideas.

Improved Quality Beyond Previous Nano Banana Models

Despite its emphasis on speed, Nano Banana 2 Lite also introduces significant quality improvements over the original Nano Banana model.

Key enhancements include:

Better World Knowledge

The model can create more accurate contextual scenes, visualisations, and location-specific content.

Stronger Character Consistency

Characters and objects remain more consistent across multiple generations, supporting storyboarding and virtual try-on experiences.

Faster Text Rendering and Localisation

The model can generate readable text directly within images, helping teams quickly produce localised advertising and marketing assets.

This combination of speed and capability makes Nano Banana 2 Lite particularly attractive for businesses operating at scale.

Creative Industry Adoption Continues to Grow

Several companies have already reported positive experiences using Nano Banana 2 Lite.

Artlist

Artlist praised the model's ability to keep creators in the creative flow without waiting on long rendering times.

Figma

Figma highlighted the model's value for rapid image exploration and creative iteration within Figma Weave.

Manus AI

Manus AI has been testing Nano Banana 2 Lite for real-time image generation across presentations, web pages, and autonomous workflows.

The common theme across these testimonials is speed without major compromises in image quality.

Enterprise Safety and Governance Built In

As AI-generated media becomes more widespread, transparency and content authenticity remain critical concerns.

Google has enabled two important safeguards by default:

C2PA Content Credentials

These credentials provide metadata that helps verify the origin and authenticity of generated content.

SynthID Watermarking

Google's imperceptible SynthID watermark technology is automatically embedded into generated outputs to help identify AI-created media.

For enterprise customers operating at scale, Google is also offering Provisioned Throughput (PT) for Nano Banana 2 Lite, ensuring reliable performance during high-volume API usage. Similar support for Gemini Omni Flash is expected soon.

The Bigger Picture

The launch of Gemini Omni Flash and Nano Banana 2 Lite reflects the next phase of generative AI adoption.

Rather than focusing solely on model quality, organisations now demand solutions that combine:

• Fast response times

• Predictable costs

• Enterprise governance

• Scalable infrastructure

• Multimodal creativity

By bringing together image generation, video editing, conversational workflows, and enterprise safeguards, Google Cloud is positioning the Gemini Enterprise Agent Platform as a comprehensive foundation for next-generation creative applications.

As businesses increasingly embed generative AI into everyday workflows, tools that balance quality, speed, and affordability are likely to become major differentiators in the AI market.