Supported Models

AgentRouter supports models from multiple mainstream LLM providers. Below is the complete list of models and pricing information.

Automatic Routing Rules

AgentRouter automatically routes to the corresponding provider based on model name:

OpenAI Models

GPT-5 Series

GPT-4 Series

GPT-3.5 Series

Usage Example

from openai import OpenAI

client = OpenAI(
    api_key="sk-ar-your-api-key",
    base_url="https://your-agentrouter.com/v1"
)

# Use GPT-4o
response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Hello!"}]
)

Anthropic Claude Models

Claude 4 Series

Claude 3.7 Series

Claude 3.5 Series

Claude 3 Series

Usage Example

from anthropic import Anthropic

client = Anthropic(
    api_key="sk-ar-your-api-key",
    base_url="https://your-agentrouter.com"
)

# Use Claude 3.5 Sonnet
message = client.messages.create(
    model="claude-3-5-sonnet-20241022",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Hello!"}]
)

DeepSeek Models

DeepSeek V3 Series

DeepSeek R1 Series

Usage Example

from openai import OpenAI

client = OpenAI(
    api_key="sk-ar-your-api-key",
    base_url="https://your-agentrouter.com/v1"
)

# Use DeepSeek
response = client.chat.completions.create(
    model="deepseek-chat",
    messages=[{"role": "user", "content": "Hello!"}]
)

Zhipu AI GLM Models

Usage Example

from openai import OpenAI

client = OpenAI(
    api_key="sk-ar-your-api-key",
    base_url="https://your-agentrouter.com/v1"
)

# Use GLM
response = client.chat.completions.create(
    model="glm-4.5",
    messages=[{"role": "user", "content": "Hello!"}]
)

Moonshot

Moonshot provides Kimi series models, using OpenAI-compatible format:

from openai import OpenAI

client = OpenAI(
    api_key="sk-ar-your-api-key",
    base_url="https://your-agentrouter.com/v1"
)

# Use Moonshot
response = client.chat.completions.create(
    model="moonshot-v1-8k",
    messages=[{"role": "user", "content": "Hello!"}]
)

Note: For specific Moonshot model pricing, please contact Moonshot official.

Pricing Explanation

Pricing Unit

All prices are in USD/M tokens (US Dollars per Million tokens).

Example:

• gpt-4o: Input $2.50/M, Output $10.00/M
• Using 1000 input tokens + 2000 output tokens
• Cost = (1000/1,000,000) × $2.50 + (2000/1,000,000) × $10.00
• Cost = $0.0025 + $0.02 = $0.0225

Billing Method

1. Pay-as-you-go - Pay only for actual tokens consumed
2. Deducted from wallet - Each request automatically deducts from wallet balance
3. Insufficient balance - Returns 402 Payment Required error

Cost Inquiry

On the Wallet page you can view:

• Current balance
• Transaction history
• Detailed cost per request
• Consumption statistics by model

Model Selection Guide

By Use Case

Daily Conversation:

• Cost-effective: gpt-3.5-turbo, deepseek-chat, claude-3-haiku
• High quality: gpt-4o, claude-3-5-sonnet

Code Generation:

• Recommended: gpt-4o, claude-3-5-sonnet, deepseek-chat

Long Text Analysis:

• Recommended: claude-3-5-sonnet (200K context)
• Alternative: gpt-4o (128K context)

Complex Reasoning:

• Recommended: gpt-5, claude-opus-4, deepseek-r1

Fast Response:

• Recommended: claude-3-haiku, gpt-4o-mini, deepseek-chat

By Budget

Low Cost (< $1/M):

• deepseek-chat: $0.14/M (input)
• claude-3-haiku: $0.25/M (input)
• gpt-4o-mini: $0.15/M (input)

Medium Cost ($1-5/M):

• gpt-4o: $2.50/M (input)
• claude-3-5-sonnet: $3.00/M (input)

Premium Models (> $10/M):

• gpt-5: $30.00/M (input)
• claude-opus-4: $15.00/M (input)

Adding New Models

AgentRouter continuously adds new model support. If the model you need is not in the list:

1. Submit a GitHub Issue
2. Contact technical support
3. For self-hosted deployments, modify lib/model-pricing.ts to add new models

Next Steps

• Getting Started - Start using these models
• OpenAI Integration Guide - Using OpenAI format models
• Anthropic Integration Guide - Using Anthropic format models
• Wallet & Billing - Learn about billing details