Claude Base Models with Smart Routing: Higher Rate Limits Across Multi-Cloud Providers

Date: 2025-11-28 / (1404-09-07)

Summary

AvalAI introduces Claude base model namespaces with smart routing across all official Anthropic cloud providers. Users can now access Claude models through simplified model names (claude-opus-4-5, claude-sonnet-4-5, claude-opus-4-1, claude-haiku-4-5) with significantly higher rate limits thanks to intelligent load distribution across Anthropic (Claude.ai), AWS Bedrock, Google Cloud Platform, and Microsoft Azure.

---

Details

Smart Routing for Claude Models

We announce the availability of Claude base model namespaces with smart routing capabilities. This feature automatically distributes API requests across all official Anthropic cloud providers, delivering significantly higher throughput and rate limits compared to single-provider access.

Supported Providers:

1. Anthropic (Claude.ai) - Direct Anthropic API
2. AWS Bedrock - Amazon Web Services
3. Google Cloud Platform (GCP) - Vertex AI
4. Microsoft Azure - Azure AI

New Base Model Names

The following base model namespaces are now available:

Base Model	AWS Bedrock Model ID	Context Window
claude-opus-4-5	anthropic.claude-opus-4-5-20251101-v1:0	200K tokens
claude-sonnet-4-5	anthropic.claude-sonnet-4-5-20250929-v1:0	200K tokens
claude-opus-4-1	anthropic.claude-opus-4-1-20250805-v1:0	200K tokens
claude-haiku-4-5	anthropic.claude-haiku-4-5-20251001-v1:0	200K tokens

Rate Limit Improvements

Smart routing enables dramatically higher rate limits by distributing load across multiple cloud providers. Here's a comparison for Tier 1 accounts:

Tier 1 Rate Limits Comparison:

Model	Namespace	Requests/min	Tokens/min
Claude Opus 4.5	anthropic.claude-opus-4-5-20251101-v1:0 (Bedrock)	1	40,000
Claude Opus 4.5	claude-opus-4-5 (Base)	10	30,000
Claude Sonnet 4.5	anthropic.claude-sonnet-4-5-20250929-v1:0 (Bedrock)	2	200,000
Claude Sonnet 4.5	claude-sonnet-4-5 (Base)	25	30,000
Claude Opus 4.1	anthropic.claude-opus-4-1-20250805-v1:0 (Bedrock)	1	200,000
Claude Opus 4.1	claude-opus-4-1 (Base)	5	30,000
Claude Haiku 4.5	anthropic.claude-haiku-4-5-20251001-v1:0 (Bedrock)	25	80,000
Claude Haiku 4.5	claude-haiku-4-5 (Base)	25	50,000

Tier 5 Rate Limits (Enterprise Grade):

Base Model	Requests/min	Tokens/min
claude-opus-4-5	1,500	4,000,000
claude-sonnet-4-5	1,500	4,000,000
claude-opus-4-1	1,500	4,000,000
claude-haiku-4-5	1,500	8,000,000

Pricing

Base model pricing remains identical to the original AWS Bedrock model pricing:

Model	Input	Cached Input	Output
claude-opus-4-5	$5.00/1M tokens	$1.50/1M tokens	$25.00/1M tokens
claude-sonnet-4-5	$3.00/1M tokens	$1.50/1M tokens	$15.00/1M tokens
claude-opus-4-1	$15.00/1M tokens	$7.50/1M tokens	$75.00/1M tokens
claude-haiku-4-5	$1.00/1M tokens	$0.50/1M tokens	$5.00/1M tokens

Migration Guide

Migrating to base model namespaces is straightforward - simply update the model name in your API calls:

Before (AWS Bedrock namespace):

model = "anthropic.claude-opus-4-5-20251101-v1:0"

After (Base model namespace):

model = "claude-opus-4-5"

No other code changes are required. The API interface, response format, and all capabilities remain identical.

API Request/Response Examples

Example Request

curl https://api.avalai.ir/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AVALAI_API_KEY" \
  -d '{
    "model": "claude-opus-4-5",
    "messages": [
      {
        "role": "user",
        "content": "Explain the benefits of multi-cloud architecture for AI workloads."
      }
    ]
  }'

Example Response

{
  "id": "chatcmpl-abc123",
  "created": 1732789200,
  "model": "claude-opus-4-5",
  "object": "chat.completion",
  "system_fingerprint": null,
  "choices": [
    {
      "finish_reason": "stop",
      "index": 0,
      "message": {
        "content": "Multi-cloud architecture for AI workloads offers several key benefits:\n\n1. **High Availability**: Distributing workloads across multiple cloud providers ensures redundancy and minimizes downtime.\n\n2. **Rate Limit Optimization**: By routing requests to multiple providers, you can achieve higher aggregate throughput than any single provider offers.\n\n3. **Cost Optimization**: Different providers may offer better pricing for specific regions or usage patterns.\n\n4. **Vendor Independence**: Avoiding lock-in to a single provider increases flexibility and negotiating leverage.\n\n5. **Geographic Coverage**: Access to data centers across different regions improves latency for global applications.",
        "role": "assistant",
        "thinking_blocks": [],
        "annotations": []
      }
    }
  ],
  "usage": {
    "completion_tokens": 150,
    "prompt_tokens": 18,
    "total_tokens": 168,
    "completion_tokens_details": null,
    "prompt_tokens_details": {
      "audio_tokens": null,
      "cached_tokens": null,
      "text_tokens": 18,
      "image_tokens": null
    }
  },
  "estimated_cost": {
    "unit": "0.0038400000",
    "irt": 440.06,
    "exchange_rate": 114600
  }
}

SDK Usage Examples

Bash

curl https://api.avalai.ir/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AVALAI_API_KEY" \
  -d '{
    "model": "claude-opus-4-5",
    "messages": [
      {
        "role": "user",
        "content": "Write a Python function for binary search."
      }
    ]
  }'

Python

from openai import OpenAI

client = OpenAI(api_key="your-avalai-api-key", base_url="https://api.avalai.ir/v1")

# Use base model namespace for smart routing
completion = client.chat.completions.create(
    model="claude-opus-4-5",  # Smart routing across all Anthropic providers
    messages=[
        {
            "role": "user",
            "content": "Write a Python function for binary search.",
        }
    ],
)

print(completion.choices[0].message.content)

JavaScript

import { OpenAI } from "openai";

const client = new OpenAI({
  apiKey: process.env.AVALAI_API_KEY,
  baseURL: "https://api.avalai.ir/v1",
});

// Use base model namespace for smart routing
const completion = await client.chat.completions.create({
  model: "claude-opus-4-5",  // Smart routing across all Anthropic providers
  messages: [
    {
      role: "user",
      content: "Write a Python function for binary search.",
    },
  ],
});

console.log(completion.choices[0].message.content);

Using Anthropic SDK

Base model namespaces are also supported with the Anthropic official SDK:

Bash

curl https://api.avalai.ir/v1/messages \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $AVALAI_API_KEY" \
  -H "anthropic-version: 2023-06-01" \
  -d '{
    "model": "claude-opus-4-5",
    "max_tokens": 1024,
    "messages": [
      {
        "role": "user",
        "content": "Explain smart routing in cloud architecture."
      }
    ]
  }'

Python

import anthropic

client = anthropic.Anthropic(
    api_key="your-avalai-api-key", base_url="https://api.avalai.ir"
)

message = client.messages.create(
    model="claude-opus-4-5",  # Smart routing enabled
    max_tokens=1024,
    messages=[
        {
            "role": "user",
            "content": "Explain smart routing in cloud architecture.",
        }
    ],
)

print(message.content[0].text)

JavaScript

import Anthropic from "@anthropic-ai/sdk";

const client = new Anthropic({
  apiKey: process.env.AVALAI_API_KEY,
  baseURL: "https://api.avalai.ir",
});

const message = await client.messages.create({
  model: "claude-opus-4-5",  // Smart routing enabled
  max_tokens: 1024,
  messages: [
    {
      role: "user",
      content: "Explain smart routing in cloud architecture.",
    },
  ],
});

console.log(message.content[0].text);

Backward Compatibility

AWS Bedrock model namespaces remain fully supported. You can continue using the full model IDs if you prefer direct Bedrock access:

• anthropic.claude-opus-4-5-20251101-v1:0
• anthropic.claude-sonnet-4-5-20250929-v1:0
• anthropic.claude-opus-4-1-20250805-v1:0
• anthropic.claude-haiku-4-5-20251001-v1:0

However, we recommend migrating to base model namespaces to benefit from higher rate limits and improved availability through smart routing.

---

Related Links

• Anthropic Models Overview
• Rate Limits Documentation
• Pricing Information
• Production Best Practices