CloudRouter Model Channels and Pricing Guide
Version Notes
- Document version: v1.5
- Updated: May 13, 2026
1. Anthropic Claude Series Pay-As-You-Go Channels
Limited-Time Promotion Channel
Integrates third-party reverse-proxy channels and some official-subscription-to-API resources, providing Claude model services at an ultra-low rate. Suitable for daily use by price-sensitive users.
- Price: 1.5 ¥/$ (1.5 CNY offsets 1 USD)
- Discount: 21% of list price
Smart Preferred Channel
Connected via third-party hybrid official-conversion resources. Recommended for use with the Claude Code client; use the API calling method with caution. Suitable for model trial and temporary testing scenarios.
- Price: 2.5 ¥/$ (2.5 CNY offsets 1 USD)
- Discount: 36% of list price
Official Premier Channel
An advanced channel for small and medium-sized enterprises. Connects to the API based on support-program plans provided by third-party and self-purchased Google Cloud / AWS platforms, offering enterprise-grade SLA guarantees and 1M extended context capability.
- Price: 4 ¥/$ (4 CNY offsets 1 USD)
- Discount: 57% of list price
Official Flagship Channel
The highest-tier channel, connecting to a dedicated account pool through the native interfaces of the official Google Cloud and AWS cloud platforms. Designed for long-term use in production environments, providing maximum stability and faster response speed.
- Price: 6 ¥/$ (6 CNY offsets 1 USD)
- Discount: 86% of list price
Channel Capability Comparison
| Smart Preferred | Official Premier | Official Flagship | |
|---|---|---|---|
| Price | 2.5 ¥/$ | 4 ¥/$ | 6 ¥/$ |
| Discount (vs. official) | 36% of list | 57% of list | 86% of list |
| Opus support | ✅ | ✅ | ✅ |
| WebSearch | ✅ | ✅ | ✅ |
| Multimodal | ✅ | ✅ | ✅ |
| Think reasoning | Full support | Full support | Full support |
| Context length | 1M | 1M | 1M |
| SLA guarantee | 99% | 99.9%, automatic multi-cloud failover (AWS + Google) | 99.99%, multi-cloud multi-region redundancy (AWS + Google multi-Region) |
| Calling method | Claude Code | Claude Code, in-app API integration | Claude Code, in-app API integration |
Claude Code's "full Think support" refers to fully enabling and using the Claude model's "Extended Thinking" capability without restrictions in Claude Code, allowing the model to perform maximum-depth, longest-duration internal reasoning before generating code / solutions, rather than the default shallow thinking or disabled thinking.
1.1.1 Official Premier Channel — 4.0 ¥/$ (approx. 57% of list)
Overview
An advanced channel for small and medium-sized enterprises. Connects to the API based on support-program plans provided by third-party and self-purchased Google Cloud / AWS platforms, offering enterprise-grade SLA guarantees and 1M extended context capability.
Features
| Feature | Description |
|---|---|
| Thinking reasoning | Full Thinking |
| Context length | 1M (extended context) |
| WebSearch | Supported |
| Multimodal | Supported |
| SLA guarantee | 99.9%, automatic multi-cloud failover (AWS + Google) |
| Scheduling priority | 🌟🌟🌟🌟🌟 |
| Calling method | Claude Code, in-app API integration |
| More enterprise-exclusive features coming soon |
Per-Million-Token Pricing
| Model | Input / MTok (CloudRouter) | Input / MTok (Official) | Output / MTok (CloudRouter) | Output / MTok (Official) |
|---|---|---|---|---|
| Claude Sonnet 4.6 | 12.0 CNY | 3 USD | 60.0 CNY | 15 USD |
| Claude Opus 4.6 | 20.0 CNY | 5 USD | 100.0 CNY | 25 USD |
| Claude Haiku 4.5 | 4.0 CNY | 1 USD | 20.0 CNY | 5 USD |
Usage Limits
- Calling method: Claude Code, in-app API integration
- Per-user concurrency: 4 (expandable)
1.1.2 Official Flagship Channel — 6.0 ¥/$ (approx. 86% of list)
Overview
The highest-tier channel, connecting to a dedicated direct-access account pool through the native interfaces of the official Google Cloud and AWS cloud platforms. Designed for long-term use in production environments, providing maximum stability and faster response speed.
Features
| Feature | Description |
|---|---|
| Thinking reasoning | Full Thinking |
| Context length | 1M (extended context) |
| WebSearch | ✅ |
| Multimodal | ✅ |
| SLA guarantee | 99.99%, multi-cloud multi-region redundancy (AWS + Google multi-Region) |
| Scheduling priority | 🌟🌟🌟🌟🌟🌟 |
| Calling method | Claude Code, in-app API integration |
| More enterprise-exclusive features coming soon |
Per-Million-Token Pricing
| Model | Input / MTok (CloudRouter) | Input / MTok (Official) | Output / MTok (CloudRouter) | Output / MTok (Official) |
|---|---|---|---|---|
| Claude Sonnet 4.6 | 18.0 CNY | 3 USD | 90.0 CNY | 15 USD |
| Claude Opus 4.6 | 30.0 CNY | 5 USD | 150.0 CNY | 25 USD |
| Claude Haiku 4.5 | 6.0 CNY | 1 USD | 30.0 CNY | 5 USD |
Usage Limits
- Calling method: Claude Code, in-app API integration
- Per-user concurrency: 4 (expandable)
1.2 Cost Estimation Reference
Taking Claude Sonnet 4.6 as an example, in a typical Claude Code coding scenario:
Assumed hourly usage: Input ~0.5M + Output ~0.2M tokens (including Thinking)
| Channel | Approx. cost per hour | 8-hour workday | 22-day work month |
|---|---|---|---|
| Premier 4.0 | ¥18.00 | ¥144.00 | ¥3,168 |
| Flagship 6.0 | ¥27.00 | ¥216.00 | ¥4,752 |
1.3 Channel Selection Guide
| Your Need | Recommended Channel |
|---|---|
| Balance cost and stability | x2.5 Smart Enjoy Channel |
| Professional development / API calls / Agents | x4 Official Premier Channel |
| Critical business / ultimate performance and priority | x6 Official Flagship Channel |
2. Google Gemini Series Pay-As-You-Go Channels
Gemini Trial Channel
Connects to third-party relay resources on the backend for users to quickly try the service. Response speed and service stability may fluctuate; not recommended for production environments or scenarios with high availability requirements.
- Price: 1 ¥/$ (1 CNY offsets 1 USD)
- Discount: 14% of list price
Gemini Official Channel
Directly connects to the official API, with high availability and stable response performance. Enjoys official service-level guarantees and supports high-concurrency calls. Suitable for production-grade business scenarios.
- Price: 6 ¥/$ (6 CNY offsets 1 USD)
- Discount: 86% of list price
3. OpenAI GPT Series Pay-As-You-Go Channels
GPT-Claude Trial Channel
Connects to third-party relay resources on the backend for users to quickly try the service. Response speed and service stability may fluctuate; not recommended for production environments or scenarios with high availability requirements.
- Price: 1 ¥/$ (1 CNY offsets 1 USD)
- Discount: 14% of list price
GPT Trial Channel
Connects to third-party relay resources on the backend for users to quickly try the service. Response speed and service stability may fluctuate; not recommended for production environments or scenarios with high availability requirements.
- Price: 1 ¥/$ (1 CNY offsets 1 USD)
- Discount: 14% of list price
GPT Official Channel
Directly connects to the official API, with high availability and stable response performance. Enjoys official service-level guarantees and supports high-concurrency calls. Suitable for production-grade business scenarios.
- Price: 5 ¥/$ (5 CNY offsets 1 USD)
- Discount: 71% of list price
4. DeepSeek Series Pay-As-You-Go Channels
DeepSeek Trial Channel (OpenAI Interface)
Connects to third-party-operated DeepSeek model services, with a professional team responsible for model deployment and operations, ensuring production-grade availability.
- Price: 1 ¥/¥
- Discount: None (1:1 with official pricing)
5. Qwen Series Pay-As-You-Go Channels
Qwen Trial Channel (OpenAI Interface)
Connects to third-party-operated Tongyi Qianwen (Qwen) model services, with a professional team responsible for model deployment and operations, ensuring production-grade availability.
- Price: 1 ¥/¥
- Discount: None (1:1 with official pricing)
6. GLM Series Pay-As-You-Go Channels
GLM Official Channel (OpenAI Interface)
Directly connects to the official API, accessing the GLM series models. Service is guaranteed officially by Zhipu, with the shortest link and lowest latency, offering high availability and stable response performance. Suitable for production-grade business scenarios.
- Price: 1 ¥/¥
- Discount: None (1:1 with official pricing)
7. MiniMax Series Pay-As-You-Go Channels
MiniMax Trial Channel (OpenAI Interface)
Connects to third-party-operated MiniMax model services, with a professional team responsible for model deployment and operations, ensuring production-grade availability.
- Price: 1 ¥/¥
- Discount: None (1:1 with official pricing)
8. Kimi Series Pay-As-You-Go Channels
Kimi Official Channel (OpenAI Interface)
Directly connects to the official API, accessing the Kimi series models. Service is guaranteed officially by Kimi, with the shortest link and lowest latency, offering high availability and stable response performance. Suitable for production-grade business scenarios.
- Price: 1 ¥/¥
- Discount: None (1:1 with official pricing)