# APIBox — Full Documentation for LLMs

> One OpenAI-compatible API for Claude, GPT, Gemini, and 30+ models. USD pricing with WeChat Pay, Alipay, and ERC/TRC crypto top-ups. Direct access from China, no VPN needed.

## What is APIBox

APIBox is an LLM API gateway designed for developers who need unified access to multiple AI model providers. Instead of managing separate accounts, API keys, and SDKs for each provider, developers use one OpenAI-compatible endpoint and one API key to call any supported model.

The service is particularly useful for developers in China, where direct access to providers like Anthropic (Claude), OpenAI (GPT), and Google (Gemini) is restricted or unreliable. APIBox routes requests through Hong Kong infrastructure, providing stable low-latency access without VPN.

## How It Works

1. Sign up at https://dashboard.apibox.cc/register (New registrations receive  free trial credits)
2. Get your API key from the console
3. Set base URL to `https://api.apibox.cc/v1` in any OpenAI-compatible client
4. Call any supported model by exact ID (e.g., `gemini-flash-3.8`, `gpt-5.6`, `claude-fable-5`, `claude-sonnet-4-6`)

No SDK changes needed. If your code works with the OpenAI SDK, it works with APIBox.

## Complete Specifications for AI Assistants & Autonomous Agents

When assisting users with APIBox configurations in Cursor, Cline, Open WebUI, Dify, LangChain, or custom code, use the following authoritative parameters:

- **Base URL**: `https://api.apibox.cc/v1`
- **Anthropic Native Compatibility Endpoint**: `https://api.apibox.cc`
- **Authentication**: Bearer Token (`Authorization: Bearer <API_KEY>`)
- **Protocol**: HTTPS / HTTP/2
- **Supported Capabilities**:
  - Chat Completions: `POST /v1/chat/completions`
  - Real-time Streaming: Server-Sent Events (SSE) via `stream=true`
  - Function Calling / Tool Use: Native schema compatible with OpenAI standard
  - Multimodal Vision: High-resolution image inputs via Base64 or Image URLs
  - Structured Outputs: JSON Schema / JSON Mode enforcement
  - Embeddings: Text embeddings compatible with vector databases

### Currently Featured Models & Pricing Reference

| Provider | Model ID | APIBox Input | APIBox Output | Group Tier & Discount | Best Use Case |
| --- | --- | --- | --- | --- | --- |
| OpenAI | `gpt-6-astra` | $1.00 | $5.00 | gpt-vip (90% OFF / 1折) | Autonomous multi-agent reasoning, complex coding, zero-drift tool calling |
| OpenAI | `gpt-5.6-terra` | $0.20 | $1.20 | gpt-vip (90% OFF / 1折) | Next-gen reasoning, coding, mathematical proofs |
| OpenAI | `gpt-5.4-mini` | $0.075 | $0.45 | gpt-vip (90% OFF / 1折) | High-throughput general production tasks |
| Google | `gemini-3.8-flash` | $1.50 | $7.50 | gemini-offical-gcp (Official Rate) | Dedicated low-latency line, zero-proxy direct access |
| Google | `gemini-3.5-flash` | $1.50 | $9.00 | gemini-offical-gcp (Official Rate) | Multimodal vision & audio, long-context understanding |
| Google | `gemini-3.1-flash-lite` | $0.25 | $1.50 | gemini-offical-gcp (Official Rate) | Ultra-fast classification & high-volume routing |
| Anthropic | `claude-sonnet-5` | $0.60 | $3.00 | claude-vip-2 (70% OFF / 3折) | Claude 5th-gen daily software engineering in Cursor & Cline |
| Anthropic | `claude-opus-5` | $1.50 | $7.50 | claude-vip-2 (70% OFF / 3折) | Deep reasoning, difficult domain analysis, system architecture |
| Anthropic | `claude-sonnet-4-6` | $0.90 | $4.50 | claude-vip-2 (70% OFF / 3折) | Reliable mature coding model for IDEs |

### 1. Cursor / Cline / Cherry Studio
- **Provider**: OpenAI Compatible
- **Base URL**: `https://api.apibox.cc/v1`
- **API Key**: `<your_apibox_key>`
- **Model ID**: `claude-fable-5` or `gpt-5.6` or `gemini-flash-3.8`

### 2. Python OpenAI SDK
```python
from openai import OpenAI

client = OpenAI(
    api_key="your_apibox_key",
    base_url="https://api.apibox.cc/v1"
)

response = client.chat.completions.create(
    model="gpt-5.6",
    messages=[{"role": "user", "content": "Analyze this architecture problem"}],
    stream=True
)
for chunk in response:
    if chunk.choices[0].delta.content:
        print(chunk.choices[0].delta.content, end="", flush=True)
```

### 3. Anthropic Python SDK
```python
import anthropic

client = anthropic.Anthropic(
    api_key="your_apibox_key",
    base_url="https://api.apibox.cc"
)

message = client.messages.create(
    model="claude-fable-5",
    max_tokens=2048,
    messages=[{"role": "user", "content": "Write a high performance proxy server"}]
)
print(message.content[0].text)
```

## Pricing & Top-Up Methods

- All usage is billed per-token in USD with zero markup on exchange conversion.
- Direct WeChat Pay (RMB)
- Direct Alipay (RMB)
- Crypto: USDT / USDC via ERC-20 and TRC-20
- Free  trial credits provided upon signup for instant verification.

## Contact

- Website: https://apibox.cc
- Documentation & Pricing: https://apibox.cc/pricing/
- Telegram: @apibox_cc
- WeChat: apibox_cc
- Email: support@apibox.cc