One key for China’s leading models

Why teams use it

Supported models

DeepSeek

Open-weight reasoning model, strong at code and math

DeepSeek API →

Qwen

Alibaba's open model family - broad and multimodal

Qwen API →

Doubao

ByteDance's model line - strong Chinese dialogue at low cost

Doubao API →

Kimi (Moonshot)

Built for very long context

Kimi (Moonshot) API →

Zhipu GLM

Tsinghua-lineage model family from Zhipu AI

Zhipu GLM API →

Hunyuan

Tencent's in-house model behind Yuanbao

Hunyuan API →

MiniMax

Text, speech and video in one family

MiniMax API →

Image and video generation are supported too, including Kling and Seedance. Current availability is listed on the pricing page.

Who it fits

• Building an AI app or agent that needs model fallbacks
• Solo developers tired of one account and one top-up per vendor
• Teams that need auditable, traceable API billing

Get started

1. Create an account in the console.
2. Top up inside WeChat Pay.
3. Issue an API key and point your client at the gateway base URL.
4. Estimate spend first with the token calculator.

Full request and parameter reference lives in the developer docs.

FAQ

What is FC SAS Token?

An AI API gateway operated by Fengchao Surges. One key reaches many leading models, metered in tokens and billed inside WeChat Pay.

Is it OpenAI compatible?

Yes. Requests follow the OpenAI chat completions shape, so you can keep your existing SDK and swap the base URL.

Which models are supported?

Text models including DeepSeek, Qwen, Doubao, Kimi, GLM, Hunyuan and MiniMax, plus image and video generation such as Kling and Seedance. See the models page for the current list.

How does billing work?

Usage is metered in tokens and settled from a single balance topped up inside WeChat Pay. Published rates are listed on the pricing page.

Who is it for?

Developers building in China who need domestic compliance, one reconciliation trail, and model fallbacks without signing up to each vendor separately.