Model API Service
Gogpu Cloud Model API · OpenAI / Anthropic compatible · Token packages + pay-as-you-go
Use the Model API to call large language models from your apps, scripts, or tools like Cursor without buying a GPU VM. Create an API key in the console and set the base URL to https://mass.gogpu.cn.
TIP
New to LLMs and Token billing? Read Quick Start first, then Billing and API Guide.
Core capabilities
| Capability | Description |
|---|---|
| Chat & text | Q&A, writing, summarization, code assist (models listed in console) |
| Tool compatibility | OpenAI-style and Anthropic-style HTTP APIs — change base URL + key in most clients |
| Billing | Token packages (prepaid quota) or pay-as-you-go (balance deducted by usage) |
| Usage | Usage stats in console; package orders and balance details for billing |
| Keys | Create and manage sk- keys; set daily/monthly spend caps (CNY) for pay-as-you-go keys |
API endpoints
| Purpose | URL |
|---|---|
| Chat (OpenAI style) | https://mass.gogpu.cn/v1/chat/completions |
| Chat (Anthropic style) | https://mass.gogpu.cn/anthropic/v1/messages |
| List models | GET https://mass.gogpu.cn/v1/models (requires API key) |
Base URL for tools:
- OpenAI-compatible clients:
https://mass.gogpu.cn/v1 - Anthropic-compatible clients:
https://mass.gogpu.cn/anthropic
WARNING
https://mass.gogpu.cn is an API endpoint, not a website. Configure it in code or AI tools; do not expect a page in the browser.
Model API vs GPU model marketplace
| Console · Model API | GPU model marketplace (platform) | |
|---|---|---|
| Where | Console → Model API → model list | Platform GPU docs (Chinese) |
| Use case | Call hosted models via HTTP API | Run open-source models on your GPU VM |
| Billing | Token packages or pay-as-you-go | Mainly GPU instance hours |
Next steps
| Goal | Doc |
|---|---|
| First API call in ~5 min | Quick Start |
| curl, Python, errors | API Guide |
| Pricing & settlement | Billing |
| Troubleshooting | FAQ |
Full Chinese documentation: 简体中文首页