Blog
Nano Banana 2 Lite:Google史上最速・最安のAI画像モデルNEW RELEASE
2026年7月1日7分で読めます

Nano Banana 2 Lite:Google史上最速・最安のAI画像モデル

On June 30, 2026, Google quietly reshuffled its image-generation lineup with Nano Banana 2 Lite — a stripped-down, speed-first sibling to Nano Banana 2 and Nano Banana Pro. Its pitch is simple: give developers and creators a model built for volume, not for the single perfect hero shot.

The headline numbers are hard to ignore. A text-to-image generation lands in about 4 seconds, priced at roughly $0.034 per 1K-resolution image — a fraction of what Nano Banana 2 or Pro cost at higher resolutions. Google shipped it alongside Gemini Omni Flash, a companion video model, so the same cheap image can be animated right after it's generated.

Note: Nano Banana 2 Lite is not yet available on HeyMarmot. HeyMarmot already offers Nano Banana 2 and Nano Banana Pro for image generation — we're evaluating Lite for high-volume, rapid-iteration workflows and will update this post when it lands.

What You'll Learn

  • What Nano Banana 2 Lite actually is and how it differs from the rest of the Nano Banana family
  • Real pricing and speed numbers, not marketing rounding
  • Where it's available — API, AI Studio, Enterprise, and consumer surfaces
  • The trade-offs you take on for that speed and price
  • What this means if you create images or video at scale

The Headline: 4-Second Images at a Fraction of the Cost

Nano Banana 2 Lite is built around one idea — throughput over ceiling. Google describes it as the fastest and most cost-efficient model in the Nano Banana family, optimized for creators who need to "workshop images and produce a large number of them in quick succession" rather than chase maximum creative control on every single frame.

In practice that means:

  • ~4 seconds per text-to-image generation
  • ~$0.034 per 1K-resolution image
  • 1K resolution only — no 2K or 4K output, unlike its bigger siblings
  • API identifier gemini-3.1-flash-lite-image, positioned as the direct replacement for the original (now legacy) Nano Banana / gemini-2.5-flash-image

For comparison, Nano Banana 2's standard tier runs roughly $0.151 per 4K image, and Nano Banana Pro around $0.24 per 4K image. Lite isn't a cheaper way to get the same output — it's a different tool for a different job.

Where It's Available

Google rolled Nano Banana 2 Lite out broadly on day one:

  • Developers: Google AI Studio, the Gemini API, and the Gemini Enterprise Agent Platform
  • Consumers: AI Mode in Search, the Gemini app, NotebookLM, Google Photos, Google Stitch, Google Flow, and Google Ads

That's a wider simultaneous rollout than most Gemini image model launches, which signals Google wants Lite to become the default for anything that doesn't need Pro-level fidelity.

How It Fits in the Nano Banana Family

Nano Banana 2 LiteNano Banana 2Nano Banana Pro
API namegemini-3.1-flash-lite-imagegemini-3-pro-image familygemini-3-pro-image (Pro tier)
Resolution1K only0.5K / 1K / 2K / 4K1K / 2K / 4K
Speed~4 seconds (fastest)BalancedMore deliberate
Price (4K)Not supported~$0.151~$0.24
Text-to-image Elo12511245
Reference imagesFewer, simpler editsUp to 14Up to 14
Best forHigh-volume batch generation, rapid iterationGeneral-purpose productionComplex final renders, brand-critical output

The Elo number is the surprising bit: on Google's own text-to-image leaderboard, Lite edges out Pro (1251 vs 1245). That's not a contradiction — it reflects raw single-prompt image quality, not the deeper reasoning, multi-reference composition, and 4K fidelity that Pro is actually built for. Lite is tuned to be good enough, fast — not to out-think your prompt.

The Trade-offs

Speed and price this aggressive come with real limits:

  • No 2K/4K output. If you need print-quality or large-format assets, Lite tops out below what Nano Banana 2 and Pro deliver.
  • Less reasoning depth. Nano Banana 2 and Pro apply heavier multi-step reasoning to interpret complex prompts and multi-image references. Lite trades some of that for latency.
  • Simpler editing. Complex compositing, style transfer, and multi-reference workflows are still Nano Banana 2 / Pro territory.

Think of Lite as the model you reach for when you're generating the twentieth variation, not the first.

The Companion Release: Gemini Omni Flash

Google didn't ship Nano Banana 2 Lite alone — it launched alongside Gemini Omni Flash, a fast, low-cost image-to-video model. Together they form a pipeline: generate a still cheaply with Lite, then animate it cheaply with Omni Flash. Google's framing is direct — building generative media is "often about creative iteration," and this pairing is meant to make high-volume, end-to-end multimedia production practical at a price point that didn't really exist before.

What This Means for Creators

  • Batch and catalog work — product variations, thumbnail sets, and ad creative at scale finally have a purpose-built, affordable model
  • Rapid prototyping — rough out a composition or concept in seconds before committing a final pass to Nano Banana Pro
  • Image-to-video pipelines — pair Lite with Omni Flash to go from prompt to cheap animated clip in one workflow
  • Cost-sensitive, high-frequency apps — anything generating images programmatically at scale (search results, ad platforms, chat assistants) gets a meaningfully cheaper default

The bigger signal here is that Google is explicitly segmenting its image lineup by job, not just by quality tier: Lite for volume, Nano Banana 2 for general production, Pro for the shots that need to be right. That's a more mature product strategy than "one model, three sizes" — and it's likely the shape more AI image providers converge on next.

Conclusion

Nano Banana 2 Lite isn't trying to be the best image model Google has ever shipped — it's trying to be the cheapest, fastest one that's still good enough to use in production. At 4 seconds and $0.034 per image, it opens up workflows that were previously too slow or too expensive to run at scale.

HeyMarmot already supports Nano Banana 2 and Nano Banana Pro for image generation — you can start creating with them today in the HeyMarmot workspace. We're keeping an eye on Nano Banana 2 Lite for high-volume, rapid-iteration use cases and will update this post as soon as it's part of the lineup.