GPT, Claude, GLM, DeepSeek, Qwen, Kimi, MiniMax & more — served from a single /v1/chat/completions endpoint with per-model pricing you can audit.
Frontier models from OpenAI, Anthropic, Z.ai, DeepSeek, Qwen, Moonshot and Xiaomi on one account.
Drop-in base_url swap for Claude Code, Cursor, Cherry Studio, One API or your own app.
Alipay & WeChat auto recharge — no waiting on manual order processing.
Each model page shows measured benchmark index and exact per-token price.
| Model | Input $/M | Output $/M | Coding |
|---|---|---|---|
| codex-auto-review | $0.018 | $0.11 | — |
| deepseek-v3.2 | $0.002 | $0.003 | 44.2 |
| deepseek-v4-flash | $0.002 | $0.005 | 69.1 |
| deepseek-v4-flash-0731 | $0.002 | $0.005 | — |
| deepseek-v4-pro | $0.005 | $0.015 | 68.8 |
| deepseek-v4-pro-0813 | $0.005 | $0.015 | — |
| doubao-seed-2.0-lite | $0.001 | $0.004 | — |
| doubao-seed-2.1-turbo | $0.003 | $0.016 | — |
| doubao-seed-evolving | $0.007 | $0.033 | — |
| gemini-3.1-pro-preview | $0.014 | $0.09 | 68.8 |
| gemini-3.1-pro-preview-thinking-128 | $0.014 | $0.09 | — |
| gemini-3.6-flash | $0.011 | $0.05 | 69.2 |
pip install openai
from openai import OpenAI
client = OpenAI(
base_url="https://starseaapi.com/v1",
api_key="sk-..."
)
r = client.chat.completions.create(
model="deepseek-v4-pro",
messages=[{"role":"user",
"content":"Hello"}]
)
Works with any OpenAI SDK, Claude Code, Cursor and 20+ tools.