> ## Documentation Index
> Fetch the complete documentation index at: https://docs.laozhang.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# GPT-Image-2-VIP and Nano Banana Production Stability Update

> GPT-Image-2-VIP supports 4K and quality again, while Nano Banana Pro, Nano Banana 2, and Lite are stable for high-concurrency production traffic.

**Published: July 22, 2026**

Good news: `gpt-image-2-vip` is stably serving 4K sizes and the `quality` parameter again. Nano Banana Pro, Nano Banana 2, and Nano Banana 2 Lite are also stable for high-concurrency production traffic. Both product lines use official resources with substantial discounts or predictable per-request pricing.

## ✨ GPT-Image-2-VIP

<Card>
  **Low-cost, stable 4K image generation and editing**

  * 🖼️ **Image generation**: `POST /v1/images/generations`
  * 🪄 **Image editing**: `POST /v1/images/edits`
  * 🤖 **Model ID**: `gpt-image-2-vip`
  * 📐 **Sizes**: common 1K / 2K / 4K outputs
  * 💎 **Quality**: `low`, `medium`, or `high`
  * 💰 **Unified price**: **\$0.03/call** for generation or editing
  * 🔑 **Token requirement**: use a **Default Group per-request billing token**
</Card>

## 🚀 Nano Banana Series

<CardGroup cols={3}>
  <Card title="Nano Banana Pro" icon="banana">
    `gemini-3-pro-image`

    **\$0.09/call** for professional design, advanced editing, and 1K / 2K / 4K output.
  </Card>

  <Card title="Nano Banana 2" icon="zap">
    `gemini-3.1-flash-image`

    **\$0.055/call** for general 0.5K–4K production and high-throughput generation or editing.
  </Card>

  <Card title="Nano Banana 2 Lite" icon="gauge">
    `gemini-3.1-flash-lite-image`

    **\$0.025/call** for 1K drafts, batch assets, and low-latency high concurrency.
  </Card>
</CardGroup>

<Info>
  If the console still shows preview aliases, Nano Banana Pro and Nano Banana 2 remain compatible with `gemini-3-pro-image-preview` and `gemini-3.1-flash-image-preview`. Final availability, prices, and billing records are shown in the console.
</Info>

<Tip>
  Scale high-volume workloads in stages while monitoring 429s, 5xx errors, average latency, and successful-image rate. Effective throughput still depends on token group, upstream capacity, and content-safety policies.
</Tip>

## Developer Documentation

* [GPT Image 2 API](/en/api-capabilities/gpt-image-2)
* [Nano Banana Pro API](/en/api-capabilities/nano-banana-pro-image)
* [Nano Banana 2 API](/en/api-capabilities/nano-banana2-image)
* [Nano Banana 2 Lite API](/en/api-capabilities/nano-banana-2-lite-api)
* [Open the LaoZhang API console](https://api.laozhang.ai/)