A single platform with stable Claude API integration.
No overseas credit card required, stable access to Claude API (claude-opus-5 / claude-sonnet-5), pay-as-you-go per token, streaming output, multi-channel failover.
Claude / OpenAI / Gemini Model Recommendation
Popular Model Pricing (CNY / million tokens) covering Claude, GPT, and Gemini groups. Click to view full parameters and pricing for all channels.
gemini-3.7-flash
上下文 1M
Cache hit ¥0.8
gemini-3.5-flash-lite
上下文 1 m
Cache hit ¥0.15
原生gemini-3.8-flash
上下文 1m
Cache hit ¥1.5
gemini-3.6-flash
上下文 1 M
Cache hit ¥1.3
gemini-3.5-flash
上下文 1 M
Cache hit ¥1.2
gpt-5.6-terra
上下文 1.1 m
Cache hit ¥1
Core competencies
Claude API relay: native interface, multi-channel failover, pay-as-you-go, streaming output, stable and efficient.
Native Interface
Fully compatible with Anthropic's native /v1/messages API, zero-cost switching, supporting streaming output and multi-model calls.
Multi-channel disaster recovery
Automatically switch upstream channels by priority (P1→P2); a single point of failure won't affect usage, which is key to stability.
Pay-as-you-go
Billing is based on token usage, with separate charges for input, output, and cache, ensuring transparent accounting and real-time usage monitoring via the console.
Streaming output
Native SSE streaming support with real-time responses, providing smooth experience for chat and code completion.
Data Security
End-to-end encrypted transmission for every call, with detailed usage logs for traceability, reconciliation, and auditing.
Context cache
Cache reads and writes are billed separately, significantly saving costs in long-context scenarios with controlled expenses.
Compatible with tools you're familiar with
Change the Base URL and key, and with almost zero changes to the existing code, you can switch to this site.