Direct access to official Chinese AI models. No markups. No hidden fees.
Drop-in replacement for OpenAI SDK. Just change the base URL.
Servers in Silicon Valley, California. Low latency worldwide.
No monthly fees. No minimum commitments. Pay only for what you use.
Direct API access from DeepSeek and Alibaba Cloud Bailian. Fully authorized.
All prices in USD per 1M tokens. Output is 2x the input price.
| Model | Description | Input / 1M | Output / 1M |
|---|---|---|---|
| deepseek-v4-flash | Lightning fast inference | $0.50 | $1.00 |
| deepseek-v4-pro | Flagship reasoning | $1.60 | $3.20 |
| Model | Description | Input / 1M | Output / 1M |
|---|---|---|---|
| deepseek-ai/DeepSeek-R1 | Deep reasoning model | $1.60 | $3.20 |
| deepseek-ai/V3-0324 | Efficient inference | $1.00 | $2.00 |
| deepseek-ai/V3.1 | Updated inference | $1.00 | $2.00 |
| Model | Description | Input / 1M | Output / 1M |
|---|---|---|---|
| qwen-turbo | Lightweight & fast | $0.30 | $0.60 |
| qwen-plus | Balanced performance | $0.60 | $1.20 |
| qwen-max | Flagship performance | $1.20 | $2.40 |
| qwen-vl-plus | Vision-language (image input) | $1.60 | $3.20 |
| qwen3.5-35b-a3b | Lightweight 35B | $0.20 | $0.40 |
| qwen3.5-397b | Large 397B model | $1.00 | $2.00 |
| qwen3.6-flash | Fast generation | $0.40 | $0.80 |
| qwen3.6-max-preview | Latest flagship | $1.60 | $3.20 |
| qwen3.6-plus | Balanced upgrade | $1.00 | $2.00 |
| qwen3.5-plus | Enhanced performance | $0.70 | $1.40 |
| qwq-plus | Deep thinking model | $0.80 | $1.60 |
| qvq-max | Visual reasoning | $2.00 | $4.00 |
| qwen-image-2.0-pro | Image generation | $4.00 | — |
| glm-5.1 | Zhipu flagship | $2.00 | $4.00 |
| glm-5 | Zhipu next-gen | $1.40 | $2.80 |
| MiniMax-M2.5 | MiniMax flagship | $0.80 | $1.60 |
| kimi-k2.6 | Moonshot Kimi | $2.40 | $4.80 |
Minimum top-up: $1 USD · Instant credit after payment
Top Up Nowhttps://aiomapi.com/v1Python example:
from openai import OpenAI
client = OpenAI(api_key="sk-your-key", base_url="https://aiomapi.com/v1")
response = client.chat.completions.create(
model="deepseek-v4-flash",
messages=[{"role": "user", "content": "Hello!"}]
)
print(response.choices[0].message.content)