Express API is an API relay service that gives developers worldwide access to 49 Chinese AI models through a single OpenAI-compatible endpoint. Change your base_url to https://gptapi.net.cn/v1 and use any Chinese model the same way you'd use GPT-4.
DeepSeek V4 Pro/Flash, Kimi K3/K2.7/K2.6, Qwen 3.7/3.6, Doubao Pro/Lite, GLM 5.2/5.1/5, ERNIE 5.1/4.5, MiniMax M3/M2.7, Hunyuan HY3, MiMo V2.5, StepFun 3.7/3.5. Full list: gptapi.net.cn/models-list.
Per-token billing: you pay separately for input and output tokens. Prices are transparent — see gptapi.net.cn/pricing. No hidden fees. Minimum top-up: $10 via Stripe.
Yes. New accounts get $0.10 free credit — enough for ~300 conversations using DeepSeek V4 Flash.
Register at gptapi.net.cn/sign-up. Get your key instantly from the dashboard.
Yes, 100%. Any OpenAI SDK works by just changing base_url. Works with Python, Node.js, curl, LangChain, and any OpenAI client library.
Chinese labs ship top-tier models at much lower cost than Western alternatives. For example, DeepSeek V4 Pro matches GPT-4 class performance at a fraction of the price. Plus, some models like Doubao and ERNIE have uniquely strong Chinese language capabilities.
Developers building AI applications who need access to Chinese models without managing multiple API keys across different Chinese platforms. Especially useful for coding agents, multilingual apps, and cost-sensitive production workloads.