Last refreshed 2026-10-07. Next refresh: weekly.
Why use Nano Banana 2.1 on Vercel AI Gateway?
Vercel AI Gateway offers Nano Banana 2.1 with pay-as-you-go pricing at $1.50/1M input tokens. Vercel AI Gateway is a unified AI proxy providing a single OpenAI-compatible API endpoint to 275+ models from 25+ providers including Anthropic, OpenAI, Google, Meta, Mistral, DeepSeek, xAI, Alibaba, Amazon, ByteDance, Cohere, MiniMax, MoonshotAI, KwaiPilot, Black Forest Labs, Recraft, Voyage AI, NVIDIA, and more.
Compare Nano Banana 2.1 across 3 providers to find the best fit for your use caseSetup recipe
Python + curlpip install openaiexport AI_GATEWAY_API_KEY=...import os
from openai import OpenAI
client = OpenAI(
api_key=os.environ["AI_GATEWAY_API_KEY"],google/gemini-nano-banana-2.1Request example
import os
from openai import OpenAI
client = OpenAI(
api_key=os.environ["AI_GATEWAY_API_KEY"],
base_url="https://ai-gateway.vercel.sh/v1"
)
response = client.chat.completions.create(
model="google/gemini-nano-banana-2.1",
messages=[{"role": "user", "content": "Hello"}]
)
print(response.choices[0].message.content)Gotchas
- Use provider model ID "google/gemini-nano-banana-2.1", not the LLMReference slug "gemini-nano-banana-2.1".
- creator/model-name e.g. kwaipilot/kat-coder-pro-v2
- The examples expect AI_GATEWAY_API_KEY; rename it only if your application config maps the new variable.
Compare Nano Banana 2.1 Across Providers
| Provider | Input (per 1M) | Output (per 1M) |
|---|---|---|
| Google AI Studio | $1.50 | $7.50 |
| OpenRouter | $1.50 | $7.50 |
| Vercel AI Gateway | $1.50 | $7.50 |
Pricing
| Type | Price (per 1M) |
|---|---|
| Input tokens | $1.50 |
| Output tokens | $7.50 |
Capabilities
About Nano Banana 2.1
Nano Banana 2.1 (API id gemini-nano-banana-2.1, GA 6 October 2026) is Google's high-efficiency Gemini image generation and conversational editing model, succeeding Nano Banana 2 (gemini-3.1-flash-image). First-party Gemini API docs: text/image/video/PDF in, image+text out; 131,072 input / 32,768 output tokens; Thinking levels minimal/medium/high; Google Web and Image Search grounding; up to 14 reference images; 1K/2K/4K output (default 1K); Batch API supported; caching not supported. DeepMind model card (Published 6 October 2026) says it is based on Gemini 3.6 Flash.