Qwen API in Russia — a key without a foreign card, pay in USDT
The Qwen API in Russia works through RuAPI without a VPN: you get a key right after sign-up, with no Alibaba Cloud account and no foreign card. You top up in USDT from 5 USDT — it can be bought with rubles via SBP. The same key gives you qwen3.8-max and qwen3.8-flash, as well as Claude, GPT and DeepSeek.
from openai import OpenAI
client = OpenAI(
api_key="sk-...",
base_url="https://www.ruapi.ai/v1",
)
resp = client.chat.completions.create(
model="qwen3.8-flash",
messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)The regular OpenAI SDK: put your RuAPI key in place of sk-... and set base_url to https://www.ruapi.ai/v1. To switch to the flagship, replace qwen3.8-flash with qwen3.8-max — the rest of the code stays the same.
Why the Alibaba Qwen API is easier to connect from Russia through RuAPI
Paying an Alibaba Cloud bill from Russia is hard.
Alibaba Cloud accepts Visa, Mastercard, AMEX and JCB cards enabled for international payments, plus PayPal; virtual and prepaid cards are not supported. Russian Visa and Mastercard cards haven't worked abroad since 2022, Mir is not on the list, and PayPal doesn't work in Russia either. On RuAPI you pay in USDT, from 5 USDT, and you can buy it with rubles via SBP.
The official route means an account, Model Studio and regions.
To get a key directly, you create an Alibaba Cloud account, fill in its details, activate Model Studio and pick a region; each region has its own endpoint, its own keys and its own model list. On RuAPI there is one key and one address, https://www.ruapi.ai/v1, for every Qwen model.
Qwen is easy to compare with other models.
With the same key you can test qwen3.8-max against deepseek-flash, Claude or GPT on your own tasks — just change the model field. Qwen also plugs into tools with an OpenAI-compatible provider: Cline, Kilo Code, OpenCode and Qwen Code.
How to get a Qwen API key: 3 steps
The key is issued by RuAPI, not Alibaba Cloud: your Qwen requests go through RuAPI over the OpenAI-compatible protocol. You need neither an Alibaba Cloud account nor a foreign bank card.
Sign up and top up
Sign up at www.ruapi.ai with email, Google or Yandex — no phone number. Qwen models unlock after your first top-up: the $0.05 trial we usually give new accounts only covers claude-haiku-5-5, deepseek-flash and gpt-6-luna.
Top up in USDT (TRC20, BEP20, TON, Polygon) or USDC, from 5 USDT. No USDT? Buy it with rubles via SBP following docs.ruapi.ai/en/buy-usdt. Foreign-issued cards are accepted too.
Create a key
In the Console: API Keys → "Create API key". This key works with all five Qwen models and with the rest of the catalogue. You can cap a key's spending — handy for scripts and tests.
Send your first request to qwen3.8-flash
In any OpenAI SDK set base_url to https://www.ruapi.ai/v1, your RuAPI key and model qwen3.8-flash — a ready example in Python, cURL and Node.js is at the top of this page. The charge for each request (model, tokens, amount) shows up in the Console log.
According to its documentation, Qwen Code takes an OpenAI-compatible endpoint from environment variables — set them in your shell or in ~/.qwen/.env:
- OPENAI_API_KEY — your RuAPI key
- OPENAI_BASE_URL=https://www.ruapi.ai/v1
- OPENAI_MODEL=qwen3.8-max
- with all three variables set Qwen Code switches to this mode by itself; otherwise run /auth → Custom Provider → OpenAI-compatible (or set "security": {"auth": {"selectedType": "openai"}} in ~/.qwen/settings.json)
All model IDs are in the docs: docs.ruapi.ai/en/models; your first request step by step: docs.ruapi.ai/en/quickstart
Qwen API: what to put in your code
API address
base_url: https://www.ruapi.ai/v1
format: OpenAI Chat Completions
Model IDs
qwen3.8-max · qwen3.8-flash
qwen3.8-27b
qwen3.7-max · qwen3.7-plus
Payment
USDT / USDC, from 5 USDT
rubles → USDT via SBP
or a foreign-issued card
Qwen API pricing
Qwen model prices in USD per 1M tokens, from the pricing page and updated along with it. No subscription: only the tokens you use are charged to your balance.
| model | input $/M | output $/M |
|---|---|---|
| qwen3.8-flash | $0.14 cache read $0.0288 | $0.49 |
| qwen3.8-27b | $0.54 cache read $0.11 | $2.16 |
| qwen3.8-max | $2.16 cache read $0.43 | $6.46 |
| qwen3.7-max | $2.16 cache read $0.43 | $6.46 |
| qwen3.7-plus | $0.36 cache read $0.0716 >256K: $1.07 | $1.43 >256K: $4.30 |
| deepseek-flash | $0.17 cache read $0.0120 | $1.44 |
Columns: input is new input tokens, “cache read” is the part of the prompt served from the cache (cheaper than regular input), output is the model’s reply — all per 1M tokens. deepseek-flash is included for comparison: it is another low-cost model, and it already works on the trial balance.
The grey “>256K” line is Alibaba’s rate for long prompts: if a prompt goes over the threshold, the whole request is billed at that rate.
Qwen via RuAPI vs. directly on Alibaba Cloud
| Qwen via RuAPI | Alibaba Cloud Model Studio | |
|---|---|---|
| Paying from Russia | USDT or USDC from 5 USDT — USDT can be bought with rubles via SBP; a foreign-issued card works too | Visa, Mastercard, AMEX, JCB enabled for international payments, or PayPal; virtual and prepaid cards not accepted |
| Getting a key | Sign up with email, Google or Yandex — the key is right there in the Console | Alibaba Cloud account → account details → activate Model Studio → key in the chosen region |
| Models on one key | Qwen, Claude, GPT, Gemini, DeepSeek, GLM, Kimi, MiniMax | The Model Studio catalogue of the chosen region; a key from one region doesn't work in another |
Alibaba Cloud details are from the Alibaba Cloud help center: Payment FAQ (alibabacloud.com/help/en/user-center/support/payment-faq), "Get API key" and "Free quota" for Model Studio (alibabacloud.com/help/en/model-studio), checked 1 November 2026. Visa, Mastercard and PayPal suspended operations in Russia in March 2022.