DeepSeek, GLM, Qwen and more behind a single OpenAI-compatible endpoint. Pay with WeChat Pay, Alipay, PayPal or crypto. Built for developers who want frontier-class models without the friction.
Pay with WeChat Pay, Alipay, PayPal or USDT. No foreign card required.
OpenAI-compatible base URL. Switch between DeepSeek, GLM, Qwen and Kimi with a single key.
Chinese open-weight models deliver frontier-class results at a fraction of the price of US APIs.
Per-key quotas, spend caps and transparent per-token billing in the console.
Edge relay node in Hong Kong keeps latency low for global developers.
Health monitoring and automatic failover across upstream providers.
| Model | Context | Input / 1M tokens | Output / 1M tokens |
|---|---|---|---|
DeepSeek Chat deepseek-chat | 1M | $0.30 | $0.90 |
DeepSeek Reasoner deepseek-reasoner | 1M | $0.75 | $2.98 |
Prices are examples — final rates shown in the console. Peak/off-peak and cache-hit pricing apply on select models.
DeepSeek, GLM (Zhipu), Qwen (Alibaba) and Kimi (Moonshot) families, with more added regularly.
WeChat Pay, Alipay, PayPal or USDT. See Pricing for details.
Yes — use the OpenAI SDK with our base URL and your key. Anthropic-format endpoints are also available on select models.
No. We do not store conversation content. See our Privacy Policy.
After payment, we verify and credit your quota — usually within 30 minutes during business hours.
Business plan includes invoicing and a DPA (data processing agreement).
Per-token billing: input and output are billed separately; cache-hit input is much cheaper. See the console for details.
Unused prepaid balance is refunded pro rata, except where required otherwise by law.
Enterprise customers can discuss private endpoints, dedicated capacity or self-hosted options.
Yes — pay with WeChat Pay or Alipay; the gateway node is in Hong Kong for low latency.