Google has released Nano Banana 2 Lite, the latest and most affordable model in its Nano Banana family of AI image generators. The model is now available to developers through Google AI Studio, the Gemini API, and the Gemini Enterprise Agent Platform, and is simultaneously rolling out across Google’s consumer products including the Gemini app, Google Photos, NotebookLM, and AI Mode in Search.
What Nano Banana 2 Lite Delivers
Built on Google’s Gemini 3.1 Flash Lite architecture, Nano Banana 2 Lite generates 1K-resolution images in approximately four seconds. This represents a 2.7x speed improvement over the standard Gemini 3.1 Flash Image model that powers the full Nano Banana 2. Google is pricing the model at $0.034 per 1,000 images, making it the least expensive option in the company’s creative AI lineup.
The pricing undercuts the older Nano Banana 1 at $0.039 per thousand images and represents a significant reduction from Nano Banana 2 at $0.067 and Nano Banana Pro at $0.134. According to Google’s internal assessments, the model delivers roughly 60 to 70 percent of the general capability of the standard and premium tiers while executing at substantially higher speeds and lower costs.
Despite the cost reduction, benchmark performance is strong. The model achieved a Text-to-Image Elo score of 1251 in standardised evaluations, surpassing both the legacy Nano Banana 1 at 1151 and the premium Nano Banana Pro at 1245. For editing tasks, it scored 1308 on single-image editing and 1294 on multi-image editing benchmarks. The model supports 14 aspect ratios and maintains reliable prompt adherence, character consistency, and legible text rendering within generated images.
Simultaneous Developer and Consumer Rollout
Google has adopted a dual deployment strategy for Nano Banana 2 Lite, making it available through developer APIs and consumer products simultaneously.
On the developer side, the model is accessible through Google AI Studio for experimentation and prototyping, the Gemini API for production integration, and the Gemini Enterprise Agent Platform for organisations requiring enterprise-grade reliability. Provisioned throughput is available on the Enterprise Agent Platform for teams managing high-concurrency API workloads.
On the consumer side, Nano Banana 2 Lite is now powering image generation features across several Google products. The Gemini app uses it for visual responses in conversations. Google Photos employs it for creative editing capabilities. NotebookLM has integrated it for a new Short Video Overviews feature, which generates 60-second portrait videos with AI-created animations and narrative explanations. AI Mode in Google Search now uses the model to provide visual answers to queries.
This simultaneous rollout reflects Google’s broader strategy of using consumer surfaces for large-scale user feedback while building out the developer API ecosystem for third-party adoption and enterprise integration.
Industry Partners and Early Adoption
Several major technology and creative companies have committed to integrating Nano Banana 2 Lite into their platforms and workflows.
Adobe has announced plans to bring the model to Firefly, its creative AI studio, offering users a speed-optimised option alongside Adobe’s own generation tools. The company described the integration as part of its strategy to deliver professional-grade creative AI tools with flexibility and control over how creators bring ideas to life.
WPP, one of the world’s largest advertising networks, received early access to the model and has integrated it into WPP Open, its agentic marketing platform. Teams have used it for asset localisation, product swaps, and dynamic style transfers for clients. WPP’s Chief Innovation Officer described the capabilities as representing a significant advancement for controlled AI production.
Manus AI has been testing Nano Banana 2 Lite to power real-time image generation within its autonomous agent workflows, including slide deck and web page creation. The company noted that the model’s speed allows its AI agents to iterate on visuals quickly and deliver results in seconds, with image quality approaching that of the full Nano Banana 2.
Figma has integrated the model into its Weave node-based design canvas for rapid iteration, describing it as ideal for exploring ideas while staying in the creative flow. Artlist is deploying it across its content platform to enable near-instant visual generation for its community of creators.
The Gemini Omni Flash Connection
Nano Banana 2 Lite launched alongside the public preview of Gemini Omni Flash, Google’s multimodal video generation and editing model. While the two models serve different primary functions, they are designed to work together in chained workflows.
The intended pipeline uses Nano Banana 2 Lite for rapid image generation, then passes the generated image as a reference input to Gemini Omni Flash for video creation and conversational editing. The Interactions API maintains session history and context across sequential edits, supporting up to three consecutive modifications.
Google has released several demonstration applications showcasing this pipeline. One called Anywhere allows users to upload a photo and uses Nano Banana 2 Lite to generate personalised postcards showing the user at iconic global landmarks, presented on an interactive 3D globe. Other demos illustrate how the two models can be combined to create short marketing videos and animated content from text prompts.
Gemini Omni Flash is available in public preview through Google AI Studio and the Gemini API, with provisioned throughput rollout planned in the near future. The combination of both models enables developers to build comprehensive multimedia experiences connecting image generation with video creation within a single API ecosystem.
Content Authenticity Measures
Every image generated by Nano Banana 2 Lite includes two layers of content provenance information. SynthID, Google’s invisible watermarking technology, embeds an imperceptible digital signature that can be detected programmatically to verify that an image was AI-generated. C2PA content credentials provide a standardised metadata framework for tracking content origin and modification history.
Both features are enabled by default and cannot be disabled by developers or end users. This approach aligns with growing industry efforts to ensure that AI-generated content is identifiable and that provenance information travels with the content through distribution and republishing.
Market Context and Competitive Positioning
The release of Nano Banana 2 Lite comes during a period of intense competition in the AI image generation market. Multiple companies including OpenAI, Midjourney, Black Forest Labs, and Stability AI are actively developing and releasing image generation models with varying emphasis on quality, speed, cost, and openness.
Google’s strategy with Nano Banana 2 Lite differs from the open-weight approach taken by some competitors. The model remains tightly integrated into Google’s managed cloud stack, requiring usage through Google’s APIs and cloud infrastructure. This eliminates the operational complexity of self-hosting for users but binds them to Google’s pricing terms and service availability.
The pricing is notably aggressive compared to the broader market. At $0.034 per thousand images, Nano Banana 2 Lite establishes a new benchmark for cost-efficient AI image generation at competitive quality levels. For enterprise buyers evaluating options, the combination of low per-image cost, fast generation, built-in content authenticity, and integration with Google’s broader AI ecosystem presents a compelling value proposition.
The release also signals Google’s strategic positioning of image generation as foundational infrastructure rather than a premium capability. By making high-quality image generation fast and cheap, Google is encouraging developers to build applications that treat visual content generation as a standard component rather than a specialised feature.
Technical Specifications Summary
Nano Banana 2 Lite is technically designated as Gemini 3.1 Flash-Lite Image in Google’s API documentation. Key specifications include 1K-resolution output across 14 aspect ratios, approximately four-second generation time, a unified API supporting text-to-image, editing, and multi-image composition, support for up to three sequential edits via the Interactions API, default SynthID watermarking and C2PA content credentials, and availability through Google AI Studio, Gemini API, and the Gemini Enterprise Agent Platform.
The model is available for immediate use through all listed channels, with provisioned throughput options available for enterprise deployments requiring guaranteed performance.
