Nano Banana 2 (Gemini 3.1 Flash Image Preview)
By Google. Unknown parameters, a context window of 65,536 tokens, licence Unknown.
Facts
- Released
- 2026-02-26 the release date models.dev records models.dev, read
- Licence
- Unknown
- Open weights
- No models.dev records whether the weights are open models.dev, read
- Parameters
- Unknown
- Active parameters
- Unknown
- Knowledge cutoff
- 2025-01 the knowledge cutoff models.dev records models.dev, read
- Context window
- 65,536 tokens the context length OpenRouter lists OpenRouter, read
- 65,536 tokens the context limit models.dev records for the maker's own API models.dev, read
- Longest output
- 58,982 tokens the largest output OpenRouter's first provider allows OpenRouter, read
- 65,536 tokens the output limit models.dev records for the maker's own API models.dev, read
- Tool calling
- No OpenRouter does not list tools among the supported parameters OpenRouter, read
- No models.dev records tool_call as false models.dev, read
- Structured output
- Yes OpenRouter lists structured outputs among the supported parameters OpenRouter, read
- Reasoning controls
- Yes OpenRouter lists reasoning controls, with the efforts high, minimal OpenRouter, read
- Yes models.dev records reasoning as true models.dev, read
- Inputs and outputs
- image, text in, image, text out OpenRouter, read
- Good for
- Reasoning: OpenRouter lists reasoning controls for it OpenRouter, read
- Vision: OpenRouter lists images among its inputs OpenRouter, read
Prices
| Provider | Input | Output | Context | As of | Source |
|---|---|---|---|---|---|
| Google AI Studio through OpenRouter | 0.500 USD | 3.00 USD | 65,536 | OpenRouter endpoints, read | |
| Google direct | 0.500 USD | 60.00 USD | 65,536 | older than 30 days; check the provider | models.dev, read |
Run it on your own hardware
No source lists weight files or a parameter count for this model, so its memory needs are Unknown.
Use it from your harness
Set up for Google Gemini API with the model gemini-3.1-flash-image-preview. Each endpoint page has the same setup for its own address.
OpenCode
Put this in opencode.json in your project folder:
{
"$schema": "https://opencode.ai/config.json",
"provider": {
"google": {
"npm": "@ai-sdk/openai-compatible",
"name": "Google Gemini API",
"options": {
"baseURL": "https://generativelanguage.googleapis.com/v1beta/openai",
"apiKey": "{env:GEMINI_API_KEY}"
},
"models": {
"gemini-3.1-flash-image-preview": {
"name": "Nano Banana 2 (Gemini 3.1 Flash Image Preview)",
"limit": {
"context": 65536,
"output": 65536
}
}
}
}
}
}- OpenCode reads any OpenAI-compatible address through the @ai-sdk/openai-compatible package, and an address that speaks the Responses API through @ai-sdk/openai.
From OpenCode documentation, read .
Pi
Put this in ~/.pi/agent/models.json:
{
"providers": {
"google": {
"baseUrl": "https://generativelanguage.googleapis.com/v1beta/openai",
"api": "openai-completions",
"apiKey": "$GEMINI_API_KEY",
"models": [
{
"id": "gemini-3.1-flash-image-preview"
}
]
}
}
}- The apiKey field can name an environment variable as $NAME.
From Pi documentation, read .
Codex
Codex speaks only the Responses API, and Google Gemini API documents no Responses address. A gateway that offers one can sit in between.
- Codex speaks the Responses API only: responses is the one supported wire API of a custom provider. Ollama and LM Studio are built in and start with --oss.
From Codex documentation, read .
Claude Code
Claude Code sends Anthropic Messages requests, and Google Gemini API documents no such address. A gateway that translates to that API can sit in between.
- Claude Code sends Anthropic Messages requests to ANTHROPIC_BASE_URL. Anthropic says it does not support routing Claude Code to models other than Claude through any gateway, so some features may not work with another model.
From Claude Code documentation, read .
Other ways to reach it
- OpenRouter, model
google/gemini-3.1-flash-image-preview
Sources of this page
Paid links
No link in this directory is a paid link or an ad, and no listing is paid for. The order and the contents of every list come from the sources named on this page.