Already using the OpenAI SDK? Switch to Chinese LLMs with just 2 lines of code. Same SDK, same format, 80% cheaper. Access DeepSeek, Qwen, GLM, Kimi, and 34 more models.
Before (OpenAI):
client = openai.OpenAI(
api_key="sk-...",
)After (AI Token Gateway):
client = openai.OpenAI(
api_key="your-aitoken-key",
base_url="https://eaf9553505eeb8f5-115-190-107-107.serveousercontent.com/v1"
)That's it. Every function call, every chat completion, every embedding works the same way.
| Feature | OpenAI SDK | LangChain | LlamaIndex | LiteLLM | Claude Code |
|---|---|---|---|---|---|
| Chat Completions | ✓ | ✓ | ✓ | ✓ | ✓ |
| Function Calling | ✓ | ✓ | ✓ | ✓ | ✓ |
| Streaming | ✓ | ✓ | ✓ | ✓ | ✓ |
| JSON Mode | ✓ | ✓ | ✓ | ✓ | ✓ |
| Usage | OpenAI GPT-4o | AI Token (DeepSeek-V4) | You Save |
|---|---|---|---|
| 1M input tokens | $2.50 | $0.12 | 95% |
| 1M output tokens | $10.00 | $0.20 | 98% |
| 10M tokens/day | $125/day | $3.2/day | 97% |
| Monthly (30 days) | $3,750 | $96 | 97% |
Switch between models with one parameter change:
response = client.chat.completions.create(
model="deepseek-v4", # or "qwen3-flash", "glm-4-air", "kimi-k3"...
messages=[{"role": "user", "content": "Hello!"}]
)