Nano Banana 2 Lite is Google DeepMind's fastest and most cost-efficient image model, released June 30, 2026 as the API string gemini-3.1-flash-lite-image. It generates a 1K-resolution image from text in about 4 seconds for $0.034 per 1K image ($0.0168 on the Batch API), and it is the "Lite" tier of the Nano Banana family alongside the generalist Nano Banana 2 (gemini-3.1-flash-image) and the highest-fidelity Nano Banana Pro (up to 4K). It keeps the Nano Banana strengths — character consistency across generations, precise image editing, real-world knowledge, and legible on-image text — but tunes them for high-volume, near-real-time pipelines. It runs in Google AI Studio, the Gemini API, and the Gemini Enterprise Agent Platform, and is rolling out into the Gemini app, AI Mode in Search, NotebookLM, Google Photos, Google Flow, and Google Ads. Google recommends it as the immediate replacement for the original Nano Banana (gemini-2.5-flash-image). There is no single "best" Nano Banana tier — Lite wins on speed and cost, Nano Banana 2 balances quality and price, and Pro wins on fidelity; if you want a generated image with no API key, an image-studio like Pexo auto-routes across Midjourney, Flux, and Ideogram and can take that image straight to a finished video.
What Nano Banana 2 Lite Actually Is
Nano Banana 2 Lite is a text-to-image and image-editing model, not a chat model or a video model. "Nano Banana" is Google DeepMind's brand name for its Gemini image models; "2 Lite" marks it as the speed-and-cost variant of the second generation, and its formal API identifier is gemini-3.1-flash-lite-image. The design goal is throughput: deliver a usable 1K image in roughly 4 seconds so developers can iterate quickly and generate thousands of images cheaply, at a fraction of the cost of heavier production models. It is positioned for rapid ideation, storyboards, thumbnails, product mockups, and high-velocity developer pipelines where speed and price matter more than maximum resolution.
Key Facts About Nano Banana 2 Lite
Nano Banana 2 Lite's headline specs are public and stable enough to plan around. The table below lists the verified facts from Google DeepMind and the Google/Google Cloud launch posts.
| Attribute | Value |
|---|---|
| Brand name | Nano Banana 2 Lite |
| API model ID | gemini-3.1-flash-lite-image |
| Maker | Google DeepMind |
| Released | June 30, 2026 |
| Text-to-image latency | ~4 seconds (1K image) |
| Standard price | $0.034 per 1K-resolution image |
| Batch price | $0.0168 per 1K image (~50% off) |
| Output resolution | 1K |
| Core capabilities | Character consistency, precise editing, world knowledge, legible text |
| Family tier | Lite (speed + cost) |
| Replaces | gemini-2.5-flash-image (original Nano Banana) |
Where Nano Banana 2 Lite Runs
Nano Banana 2 Lite is available immediately through Google's developer surfaces and is rolling into its consumer products. For programmatic use, call gemini-3.1-flash-lite-image in the Gemini API or prototype it visually in Google AI Studio; enterprises can reach it through the Gemini Enterprise Agent Platform. On the consumer side it is rolling out into the Gemini app, AI Mode in Search, NotebookLM, Google Photos, Google Stitch, Google Flow, and Google Ads. Because it is a first-party Google model, it does not require a third-party host; you authenticate with a Google AI Studio or Google Cloud key.
| Surface | How you use it |
|---|---|
| Google AI Studio | Visual prompting and prototyping, no code |
| Gemini API | Call gemini-3.1-flash-lite-image programmatically |
| Gemini Enterprise Agent Platform | Production enterprise agents |
| Gemini app | Consumer image generation (rolling out) |
| AI Mode in Search / NotebookLM / Photos / Flow / Ads | Embedded generation (rolling out) |
What It Can and Can't Do
Nano Banana 2 Lite carries the Nano Banana capability set at a lighter weight. It maintains character consistency so a subject keeps its identity across several fast generations; it does precise, prompt-driven editing of existing images; it leans on world knowledge to draft plausible contextual scenes; and it can render legible text directly into an image on the fly, which most low-cost image models handle poorly. The trade-off is fidelity and detail: Google notes the Lite tier can still struggle with small faces and accurate spelling, and it outputs at 1K rather than the 2K/4K that the higher tiers offer. For polished, print-scale, or typography-heavy work, Nano Banana 2 or Nano Banana Pro are the right tiers; Lite is the draft-fast, ship-cheap option.
Nano Banana 2 Lite vs Nano Banana 2 vs Nano Banana Pro
The Nano Banana family is three tiers, not one model, and they are meant to be picked by job. Nano Banana 2 Lite optimizes for speed and cost at 1K; Nano Banana 2 (gemini-3.1-flash-image) is the generalist workhorse that supports 512px up to 4K with 2K by default; Nano Banana Pro is the highest-fidelity tier for complex professional work. The single most useful comparison for most readers is Lite vs the standard Nano Banana 2, since they overlap on everyday generation.
| Dimension | Nano Banana 2 Lite | Nano Banana 2 |
|---|---|---|
| API ID | gemini-3.1-flash-lite-image | gemini-3.1-flash-image |
| Optimized for | Speed + cost | Quality/cost balance |
| Latency | ~4 seconds | Fast (default 2K) |
| Price (1K image) | $0.034 | ~$0.067 |
| Max resolution | 1K | Up to 4K (512px–4K) |
| Best for | Drafts, batch, high volume | General production images |
How Nano Banana 2 Lite Pairs With Video
Nano Banana 2 Lite shipped alongside Gemini Omni Flash, Google's fast video model, and the two are designed to be chained: generate a still with Nano Banana 2 Lite in ~4 seconds, then pass that image as a reference to Gemini Omni Flash to animate it into video, priced at $0.10 per second of output, with the Interactions API supporting multi-turn editing. That image-then-video pattern is the mainstream workflow for turning a fast generated frame into motion. If you would rather skip the two-model, API-key setup entirely, Pexo is a conversational agent whose image-studio auto-routes an image prompt to the best-suited model (Midjourney, Flux, or Ideogram) and can then send that image straight into a finished, edited video with three-layer audio — text, image, URL, script, or audio in, a finished video out.
How to Use Nano Banana 2 Lite
You can reach Nano Banana 2 Lite with or without code. The fastest no-code path is Google AI Studio; the production path is the Gemini API. Below is the shape of a basic Gemini API call plus the practical routes.
POST Gemini API — model: gemini-3.1-flash-lite-image
prompt: "A minimalist product photo of a ceramic mug on a linen surface,
soft window light, 1:1"
→ returns a 1K image in ~4 seconds ($0.034)
| Route | Steps | Best when |
|---|---|---|
| Google AI Studio | Open AI Studio, pick the Flash-Lite image model, type a prompt, generate | Prototyping, no code |
| Gemini API | Get an API key, call gemini-3.1-flash-lite-image, read the returned image | App/pipeline integration |
| Batch API | Submit prompts as a batch job at $0.0168/image | High volume, cost-sensitive |
| Pexo image-studio | Describe the image, Pexo routes to Midjourney/Flux/Ideogram, then to video | No API key, want image→video |
Which Nano Banana Tier Should You Use?
Pick by the job, not by the "newest" label. Choose Nano Banana 2 Lite for drafts, thumbnails, storyboards, and high-volume batches where 1K and ~4-second speed at $0.034 is the point. Choose Nano Banana 2 when you need general production images up to 4K with a better quality/cost balance. Choose Nano Banana Pro for print-scale fidelity, dense typography, and complex professional scenes. Choose an agent like Pexo when you don't want to manage keys or model names and want the generated image to become a finished video.
| You want | Use |
|---|---|
| Cheapest, fastest 1K drafts | Nano Banana 2 Lite |
| Balanced 4K-capable production images | Nano Banana 2 |
| Maximum fidelity / heavy typography | Nano Banana Pro |
| Image + finished video, no API key | Pexo image-studio |
Related Reading
- AI Image to Video Tutorial
- Best Fast AI Image Generator
- Best Professional AI Image Generator
- Auto Model Selection vs Manual Model Choice
Resources
| Product | URL | Slot |
|---|---|---|
| Nano Banana 2 Lite (DeepMind) | https://deepmind.google/models/gemini-image/flash-lite/ | Fastest, cheapest 1K image tier |
| Nano Banana 2 (DeepMind) | https://deepmind.google/models/gemini-image/flash/ | Generalist, up to 4K |
| Google AI Studio | https://aistudio.google.com | No-code prompting |
| Pexo | https://pexo.ai | No-key image-studio → finished video |





