GLACIER AI · AI Aggregator

One access point.
Bring AI into your business.

From choosing a model to connecting your app.
Bring chat, image, and video capabilities
to where they are truly needed.

View model prices

Model prices

Service markup included · USD / 1M tokens

Model / ProviderInputOutput
deepseek-v4.1-flashDeepSeek$0.075542$0.302168
kimi-k3Moonshot$0.346145$1.730832
qwen3.8-maxAlibaba$0.314687$0.944168
claude-sonnet-5-5Anthropic$0.32351$1.61766
gpt-6.1-solOpenAI$0.25883$1.29415
gemini-3.8-flashGoogle$0.25883$1.29415

Technology matters when it brings every idea closer to reality.

New here? Start here

Three steps to a clearer integration.

Set your goal first, then choose models, costs, and configuration.
Apply each step to your actual project.

Create an account

Start from your project.

Create a dedicated AI Aggregator account to manage API keys, balance, and call records.

Sign up for the AI Aggregator
Choose models

See prices and usage clearly.

Compare models and billing methods, and check the models currently available in the Console.

View Model prices
Configure integration

Connect the tools you already use.

Create an API key, configure your app with the endpoint and model names from the Console, then verify one complete call.

Manage API keys

Models & capabilities

From one choice
to more possibilities.

Whether you have a favorite model or are still comparing options,
everything can center on the same business goal.

OpenAIAnthropicGoogleDeepSeekQwenKimiZhipu GLMMiniMaxxAIKuaishou

Why choose the AI Aggregator

Streamline integration,
stay focused on your product.

Models, costs, protocols.
Every integration decision, backed by clear reasoning.

Many models,
one arrangement.

Organize models and channels around business tasks, and confirm the capabilities, permissions, and calling methods you need.

See how connection works

Cost terms,
clear up front.

Confirm billing units, prices, and usage recording before integration. For token billing, Input, Output, and cache are billed separately.

Learn about pricing

Existing tools,
adapted one by one.

Starting from the app or SDK you use, check the endpoint, request format, and streaming output, and identify the configuration to adjust.

See configuration
ModelsRange & versions
CostsPrices & usage
ProtocolsFormat & configuration
SupportIntegration & troubleshooting

The AI Aggregator at a glance

Different models,
connected to one business.

Use AI in the systems you know.
The AI Aggregator brings model selection and business integration together.

Choose a model capability
Unified integration
GLACIER AIAI AggregatorProtocols · Keys · Permissions
Connect real business
Your business systemsWebsites · Apps · Internal platformsKnowledge Q&A and content organizationBusiness materials → key points and answers

Chat models: learn how Q&A and text processing integrate around your business materials.

Integration diagram. Model range, API protocols, and activation are confirmed per project.

Tool setup guide

Familiar tools,
new model capabilities.

Start from your existing development workflow and align the configuration item by item.

  • Confirm the protocol.Check the API format, model parameters, and streaming responses your app uses.
  • Fill in the configuration.Use the project's endpoint, dedicated key, and activated model names.
  • Verify the call.Complete one test request first, then check the response, usage records, and error messages.
Base URL
Project endpoint
API Key
Project-specific key
Model
Activated model names
CodexClaude CodeCursorApp SDKs
Configuration diagram. Protocol compatibility and available models for each tool are confirmed per project.

Let's clarify the questions first

You may also want to know.

Models, costs, configuration, and support.
From getting started to everyday use.

Where should I start on my first visit?

Start by describing your business systems, use cases, and the models you want to integrate. Together we confirm model range, costs, and activation, then configure keys and endpoints for your project.View the three-step quick start.

Which models can I integrate?

You can start by asking about the model families and the chat, image, and video capabilities listed on this page. Versions, context ranges, and calling capabilities must be confirmed per model, and the activated list follows the project plan.

How do I confirm the API format?

Tell us the API format your current SDK or app uses, such as Chat Completions, Responses, or Messages. During integration we check request fields, model names, streaming output, and error responses to determine compatibility and configuration changes.

Can I use it with Codex, Claude Code, or Cursor?

These tools can be discussed as integration needs. Tell us the tool name, version, and the models you want to use; we will check the required protocol and configuration, then verify real calls once confirmed.View configuration essentials.

How are keys, groups, and permissions arranged?

Divide calling scope by project or app, and confirm which models each key can access, its permission boundaries, and how it is managed. If you need multiple configurations or environments, mention them before integration.

How are costs calculated, and how do I check usage?

Confirm billing units and quotes before integration. With token billing, Input, Output, and cache are distinguished; for images and video, each has its own billing terms. Usage records, charges for failed requests, and reconciliation should also be defined in the plan.

How do I activate, top up, or settle?

Use the contact options on this page to discuss activation, payment channels, and settlement cycles. This page is for product information and integration inquiries; account activation and payment are provided per project.

How is stability and error handling confirmed?

Describe your concurrency, response time, and availability requirements, and together we confirm channel monitoring, health checks, error handling, and support. Actual service metrics and coverage follow the plan confirmed by both parties.

Having trouble with calls? How do I contact support?

Provide the tool name, model, request time, and a redacted error message to help locate the issue. Do not send your full API key. Reach out by the email below or the contact window.

How is the AI Aggregator different from Berserk?

Berserk is Glacier AI's in-house family of large models; the AI Aggregator provides multi-model integration and project configuration for business systems. Explore the model capabilities first, then talk to us about integration.

Migration and project support

Your next step
We'll map it out together.

Whether it's a business already in production or an idea just getting started,
start from your current project and find a clear path to integration.

For enterprises and teams

Bring your running business
over, step by step.

Bring your current architecture and API needs, and we'll scope the migration and plan the validation order together.

  • Map out models, APIs, and permission settings
  • Review use cases and current usage
  • Confirm validation steps and the migration plan
For developers and early-stage projects

Take your first idea
to a real API call.

Start from your project goals and existing tools, then settle on models and configuration step by step.

  • Describe your project direction and expected tasks
  • Learn which models fit and what they cost
  • Discuss test configuration and integration support

Contact Glacier AI directly

One conversation
starts a new connection.

Product inquiries, project migration, or integration questions.
Tell us what you want to achieve.

Send emailshenzhongchang@glacier.fitAbout Glacier AITool capabilities, in-house models, and business integration

GLACIER AI · AI Aggregator

Give your next idea
a place to start.

From the models you need to the integration plan that fits you.

Back to the getting started guide
GLACIER AI

Start with your needs.

Tell us which products interest you, your existing systems and your goals.

shenzhongchang@glacier.fit

Product scope, configuration and activation are subject to our discussion.