Author's Note: Nano Banana 2 (gemini-3.1-flash-image-preview) is officially live on Google AI Studio. At just $0.0672 per image, it's 50% cheaper than Nano Banana Pro—but APIYI has managed to go even lower. Full API integration tutorial included.
Google AI Studio quietly rolled out a model marked as New—Nano Banana 2, with the model ID gemini-3.1-flash-image-preview. The model card boldly states: Pro-level visual intelligence with Flash-speed efficiency and reality-grounded generation capabilities.
The number that's getting developers most excited is the image generation pricing: just $0.0672 per image, which is exactly half the cost of Nano Banana Pro's $0.134. While the official documentation (ai.google.dev/gemini-api/docs/image-generation) hasn't been updated yet, the model is already available for model invocation in AI Studio.
APIYI's initial pricing is around $0.03 per call, which is less than 50% of the official price~
Core Value: By the end of this article, you'll know the official pricing for gemini-3.1-flash-image-preview, how to call the API, and how its costs compare to Nano Banana Pro.
gemini-3.1-flash-image-preview Integration Notes
Current Status
| Status Item | Details |
|---|---|
| AI Studio | Live, available for direct model invocation |
| Official Docs | Not yet updated (ai.google.dev/gemini-api/docs/image-generation) |
| Official Pricing Page | Model not yet listed |
| Vertex AI | Not yet confirmed |
| Free Tier | None, marked as Paid |
| Batch API | Not yet confirmed |
The fact that official docs aren't updated means gemini-3.1-flash-image-preview is currently in a "model live, docs lagging" state. This isn't unusual for Google's release cadence—models often hit AI Studio first, with docs and pricing pages following a few days later.
🎯 Integration Advice: No need to wait for the official docs. gemini-3.1-flash-image-preview is already fully functional in AI Studio. The API interface is consistent with Nano Banana Pro; you just need to swap the model ID. Calling via the APIYI apiyi.com API proxy service is even more convenient, as the platform automatically adapts to the latest models.
FAQ
Q1: Which one should I choose: gemini-3.1-flash-image-preview or gemini-3-pro-image-preview?
For the vast majority of scenarios, go with gemini-3.1-flash-image-preview. Its image pricing is 50% cheaper, text tokens are 75% cheaper, it's faster, and the image quality is officially labeled as "Pro-level." Unless you have extreme quality requirements or existing Pro-based workflows that are hard to migrate, I'd recommend switching to Nano Banana 2 right away.
Q2: $0.0672 per image—are there any other hidden costs in practice?
The $0.0672 is the output cost for image generation. You'll also need to pay for the text input tokens ($0.50/million tokens) for your prompt. Usually, a prompt is between 50-200 tokens, which costs about $0.000025-$0.0001—basically negligible. The total cost is primarily driven by the image output fee. You might even find additional discounts when calling through APIYI apiyi.com.
Q3: Is there any risk in integrating now since the official docs aren’t updated?
The risk is very low. It's already working fine in AI Studio, and the model card shows full pricing info. Google often opens AI Studio access before updating the documentation; it's just their standard release rhythm. I'd suggest doing some small-scale testing in AI Studio or via APIYI apiyi.com first, then scaling up once you're happy with the results.
Summary
Key highlights of gemini-3.1-flash-image-preview (Nano Banana 2) launching on AI Studio:
- Image generation at $0.0672/image: 50% cheaper than Nano Banana Pro, saving you $20,040 per 10,000 images monthly.
- Text tokens are 75% cheaper: $0.50/M for input and $3.00/M for output, significantly cutting costs for multimodal applications.
- Pro-level quality + Flash speed: Officially positioned as "Pro-level visual intelligence with Flash-speed efficiency."
The model is already available for direct model invocation in AI Studio, with official documentation expected to follow in a few days. Now's the perfect time to start testing.
We recommend using APIYI (apiyi.com) for quick access to gemini-3.1-flash-image-preview. The platform supports OpenAI-compatible formats, offers free test credits, and lets you use a single API key to call all models.
📚 References
-
Google AI Studio: gemini-3.1-flash-image-preview model card and API entry point
- Link:
aistudio.google.com - Note: Currently the only official entry point to invoke this model.
- Link:
-
Gemini API Image Generation Documentation: Official developer docs (pending update)
- Link:
ai.google.dev/gemini-api/docs/image-generation - Note: Currently only contains Nano Banana Pro info; Nano Banana 2 documentation will be added soon.
- Link:
-
Gemini API Pricing Page: Official pricing reference
- Link:
ai.google.dev/gemini-api/docs/pricing - Note: Compare token and image generation prices across different models.
- Link:
-
Nano Banana Wikipedia: Full technical evolution of the Nano Banana series
- Link:
en.wikipedia.org/wiki/Nano_Banana - Note: The development journey from the first generation to Nano Banana 2.
- Link:
Author: APIYI Technical Team
Technical Exchange: Feel free to share your experience with gemini-3.1-flash-image-preview in the comments. For more AI model resources, visit the APIYI documentation center at docs.apiyi.com.
