Gemini 3.1 API
Official Google Gemini 3.1 API and third-party access options
Model names • Pricing • Documentation • aiberm shared-volume access
What Is the Gemini 3.1 API?
The Gemini 3.1 API is Google DeepMind's programmatic interface for accessing Gemini 3.1 Pro, the company's most capable language model as of early 2026. It is available through Google AI Studio and Google Cloud Vertex AI. Developers can send text, images, audio, video, and documents to the model and receive structured responses.
This page covers two access paths: the official Google API and aiberm, a third-party aggregator that provides the same underlying API at a slightly lower price by pooling usage across its user base.
Official Google Gemini 3.1 API
Direct access through Google AI Studio and Vertex AI
Model Identifiers
The primary model name for Gemini 3.1 Pro is gemini-3.1-pro-preview-thinking. Google also exposes gemini-3.1-pro-preview for non-thinking mode. These identifiers are used in API requests to specify which model to call.
Official Pricing
Google charges $2.00 per 1 million input tokens and $12.00 per 1 million output tokens for Gemini 3.1 Pro. Context caching is available at $0.50 per 1 million tokens stored per hour. Pricing is subject to change; always verify at ai.google.dev.
Context Window
Gemini 3.1 Pro supports a 1,048,576-token input context window and up to 65,536 output tokens per request. This allows processing of entire codebases, long documents, or extended conversations in a single call.
Supported Input Types
The API accepts text, images (up to 900 per request), audio (up to 8.4 hours), video (up to 1 hour), and PDF documents. Multimodal inputs can be combined in a single request.
Thinking Mode
Gemini 3.1 Pro supports a thinking parameter that enables extended internal reasoning before generating a response. Three compute tiers are available: low, medium, and high. Higher tiers improve accuracy on complex tasks at the cost of increased latency and token usage.
Official Documentation
Google's official API documentation is available at ai.google.dev. It covers authentication, request format, streaming, function calling, grounding, and all supported parameters.
ai.google.devGemini 3.1 API via aiberm
Shared-volume access at slightly lower cost
How aiberm Works
aiberm is a third-party API aggregator. It purchases API capacity from Google in bulk and distributes that capacity across its user base. Because usage is pooled, the effective per-token cost is slightly lower than purchasing directly from Google. The underlying model is identical — requests are forwarded to Google's infrastructure.
What "Shared Volume" Means
aiberm's pricing model is based on aggregated usage. Users share a common pool of API capacity purchased from Google. This is similar to how cloud reserved instances work — buying in bulk reduces the unit cost. Individual requests are still processed independently and privately.
Supported Model Names on aiberm
On aiberm, Gemini 3.1 Pro is accessible using the model identifier gemini-3.1-pro-preview-thinking. The same model name used in the official Google API works on aiberm's endpoint.
aiberm Pricing for Gemini 3.1
aiberm prices Gemini 3.1 Pro below the official Google rate. Exact current pricing is listed at the aiberm pricing page. Prices may change as volume and costs fluctuate.
aiberm.com/pricing?tab=geminiaiberm API Documentation
aiberm provides documentation covering authentication, base URL, supported models, and request format. The API is compatible with the OpenAI SDK format, which means existing code using the OpenAI client library can be pointed at aiberm's endpoint with minimal changes.
aiberm.com/docs
NanoBanana Pro — Gemini Image Generation API
gemini-3-pro-image-preview via aiberm
What Is NanoBanana Pro?
NanoBanana Pro is aiberm's name for access to gemini-3-pro-image-preview, Google's image generation model built on the Gemini 3 architecture. It generates images from text prompts and supports image editing workflows.
Model Identifier
The underlying model is gemini-3-pro-image-preview. On aiberm, this is exposed under the NanoBanana Pro product name. The model accepts text descriptions and returns generated images.
Capabilities
gemini-3-pro-image-preview supports text-to-image generation, image editing with text instructions, and style transfer. It is designed for high-fidelity output with strong prompt adherence. Resolution and aspect ratio options are available through API parameters.
Pricing
NanoBanana Pro pricing is listed on the aiberm pricing page under the Gemini tab. Pricing is per image generated, with variations based on resolution.
aiberm.com/pricing?tab=geminiOfficial Google API vs aiberm — Key Differences
Both provide access to the same Gemini 3.1 models. The differences are in pricing, billing, and support.
| Aspect | Official Google API | aiberm |
|---|---|---|
| Model | gemini-3.1-pro-preview-thinking | gemini-3.1-pro-preview-thinking |
| Underlying Infrastructure | Google Cloud directly | Google Cloud via aiberm proxy |
| Input Token Price | $2.00 / 1M tokens | Below official rate (see pricing page) |
| Output Token Price | $12.00 / 1M tokens | Below official rate (see pricing page) |
| Billing | Google account / GCP | aiberm account |
| SDK Compatibility | Google AI SDK, Vertex AI SDK | OpenAI-compatible SDK |
| Image Generation | Available (gemini-3-pro-image-preview) | Available as NanoBanana Pro |
| Support | Google support channels | aiberm support |
Frequently Asked Questions
Common questions about Gemini 3.1 API access
What is the model name for Gemini 3.1 Pro in API calls?
The official model identifier is gemini-3.1-pro-preview-thinking for the thinking-enabled version. For non-thinking mode, use gemini-3.1-pro-preview. Both identifiers work on the official Google API and on aiberm.
How does aiberm's pricing compare to Google's official pricing?
aiberm prices Gemini 3.1 Pro slightly below Google's official rate of $2.00 input / $12.00 output per 1M tokens. The exact discount varies. Current prices are listed at aiberm.com/pricing?tab=gemini.
Is the model quality the same on aiberm as on the official Google API?
Yes. aiberm forwards requests to Google's infrastructure. The model weights, capabilities, and output quality are identical. The difference is only in billing and the API endpoint URL.
Can I use the OpenAI Python SDK to call Gemini 3.1 via aiberm?
Yes. aiberm's API is compatible with the OpenAI SDK format. You set the base_url to aiberm's endpoint and provide your aiberm API key. The model parameter accepts Gemini model names.
What is gemini-3-pro-image-preview?
gemini-3-pro-image-preview is Google's image generation model built on the Gemini 3 architecture. It generates images from text prompts. On aiberm, it is available under the product name NanoBanana Pro.
Does Gemini 3.1 Pro support function calling?
Yes. The official Google API supports function calling (also called tool use). You define a set of functions in the request, and the model can choose to call them with structured arguments. aiberm also supports this feature.
What is the maximum output length for Gemini 3.1 Pro?
Gemini 3.1 Pro supports up to 65,536 output tokens per request, which is significantly higher than most competing models. This is useful for generating long documents, detailed code, or extended analysis.
Where can I find aiberm's API documentation?
aiberm's documentation is available at aiberm.com/docs. It covers the base URL, authentication, supported models, and request examples.
Start Using Gemini 3.1 API
Access Gemini 3.1 Pro through the official Google API or via aiberm for shared-volume pricing.