Runway Adds Google's Nano Banana 2 Lite for 4-Second Image Generation
Runway adds Google's fastest image model to its platform, letting users generate images in seconds via the app or Agent
Takeaways
- Runway has added Nano Banana 2 Lite (Google's
gemini-3.1-flash-lite-image) to its platform for fast image generation. - Available now via the Runway image tool or by asking Runway's Agent to use it by name.
- The model generates 1K images in ~4 seconds at $0.034 per image, optimized for speed and volume over quality.
- Supports 14 aspect ratios, fast multi-turn edits, and legible in-image text rendering; capped at 1K resolution.
- Sits alongside Runway's own Gen-4 Image model, filling a speed-first niche for rapid ideation and high-volume pipelines.
- All outputs include Google's SynthID invisible watermark for AI content provenance.
Runway has added Nano Banana 2 Lite to its platform, giving users access to Google's fastest and cheapest image generation model directly inside the Runway app. The integration is live now, and you can also ask Runway's Agent to use it by name.
What Runway is actually shipping
Nano Banana 2 Lite is Google's brand name for gemini-3.1-flash-lite-image, the speed-optimized end of the Nano Banana image model family. Runway is surfacing it as a selectable model in its image tool, sitting alongside its own Gen-4 Image model. The announcement is squarely a platform addition: Runway users now have a new, faster option for image generation without leaving the Runway environment.
What makes this notable is how the two access paths work:
- Direct: Go to the Runway image tool and select Nano Banana 2 Lite from the model picker.
- Via Agent: Ask Runway's Agent to use Nano Banana 2 Lite in a prompt, and it will route the generation accordingly.
The model behind the integration
Nano Banana 2 Lite is designed as the efficiency specialist of the image generation family, targeting sub-2 second latency and significantly reduced compute costs to enable high-volume interactive use cases and real-time consumer applications. Google states it returns a text-to-image in about four seconds at $0.034 per 1K-resolution image, while keeping reliable prompt adherence, strong character consistency, and legible in-image text.
Within the Nano Banana family, Nano Banana 2 Lite (gemini-3.1-flash-lite-image) is Google's fastest and cheapest Gemini image model, engineered for velocity and scale. The tradeoff is resolution: the Gemini 3.1 Flash Lite image model only supports 1K images, while the fuller Nano Banana 2 model goes up to 4K. For most rapid-iteration workflows inside Runway, 1K is plenty.
Here is a quick look at where it sits in the family:
| Model | Speed | Max Resolution | Price per image |
|---|---|---|---|
Nano Banana 2 Lite (gemini-3.1-flash-lite-image) |
~4 sec | 1K | $0.034 |
Nano Banana 2 (gemini-3.1-flash-image) |
Higher latency | 4K | Higher |
Nano Banana Pro (gemini-3-pro-image) |
Slowest | 4K | Premium |
What it is good at inside Runway
The model is optimized for fast multi-turn local edits like swapping colors, sticker creation, and background adjustments, while maintaining high character alignment matching original Nano Banana standards. It supports a discrete set of 14 aspect ratios including standard formats, so common creative dimensions are covered.
The practical sweet spot for this model in Runway is:
- Rapid ideation and moodboarding where you need many variations fast
- Sticker and icon generation at scale
- Background swaps and quick compositional edits
- Generating a base image to then hand off to Runway's video tools for animation
The model is capable of generating legible, stylized text for infographics, menus, diagrams, and marketing assets, which is a genuine differentiator. Most image models still struggle badly with text rendering. Grounding with Google Search, which lets the model verify facts and generate imagery based on real-time data, is not supported by the Gemini 3.1 Flash Lite Image model, so don't expect it to pull in live information.
Where it fits in Runway's broader model lineup
Runway already has its own Gen-4 Image model, which is built around reference-driven consistency and art direction. Building with generative media is often about creative iteration, and with models like this, developers can build comprehensive end-to-end multimedia experiences that connect rapid image generation with video creation and editing. Nano Banana 2 Lite fills a different niche: pure speed for high-volume generation, where you want to blast through visual options quickly before committing to a direction.
The Agent integration is the more interesting detail here. Rather than manually switching models in the UI, you can instruct Runway's Agent to use Nano Banana 2 Lite as part of a larger workflow. That positions it as a building block inside agentic pipelines, not just a standalone image generator.
How to get started
The model is available now in the Runway app. Select it from the image tool's model picker, or prompt the Agent directly. All generated images include a SynthID watermark, Google's invisible provenance marker for AI-generated content.
If you want to use the underlying model directly outside of Runway, Nano Banana 2 Lite is available in Google AI Studio, the Gemini API, and the Gemini Enterprise Agent Platform. The official Gemini API docs cover the full integration details, and Google's announcement post includes prompting guides and demo apps to get you started.