Claude in Russia: how to use it in a chat, via API and in Claude Code
Anthropic does not support Russia, but Claude models are available through Zukko: a browser chat, the API and Claude Code on one key, paid in rubles. Setup and prices.
6 min · · models · payments
You can't use Claude in Russia directly: Anthropic does not support the region and does not accept Russian cards. The Claude models themselves — Opus 5.5, Sonnet 5 and others — are available through Zukko: in a browser chat, via the API and in Claude Code, paid in rubles. Below are three ways to connect, the prices, and what you need to know about caching.
Why Claude is not available from Russia directly
Russia is not on the list of countries Anthropic supports for Claude and its API (we checked the list on Anthropic's website on October 7, 2026). You can't pay for a Claude subscription or top up the Anthropic Console with a Russian bank card. As a result, both the chat and an Anthropic API key are hard to get from Russia, and an account set up against the service's rules can be blocked together with its balance.
Zukko takes a different route: it is an API gateway to Claude, GPT, Gemini, DeepSeek, Qwen and GLM models. You get one fetch_... key, a balance in US dollars and pay-per-token billing with no subscription.
Claude in Russia: three ways through Zukko
1. Browser chat
Good if you just want to write, work through texts or ask questions.
- Sign up at zukko.pro and top up your balance (from $1).
- Open the chat.
- Pick a Claude model from the list and write.
Usage is charged per token from your balance. All Claude models in the table below accept image input.
2. API
For your own scripts and apps, create a key in the API keys section. The full key is shown only once, so save it. You can call Claude models in the Anthropic format with the official SDK:
from anthropic import Anthropic
client = Anthropic(base_url="https://api.zukko.pro", api_key="fetch_your_key")
message = client.messages.create(
model="claude-sonnet-5",
max_tokens=1024,
messages=[{"role": "user", "content": "Shorten this text to three bullet points: ..."}],
)
print(message.content[0].text)Note: for the Anthropic SDK the address has no /v1; the SDK adds it itself. If your code uses the OpenAI SDK, you can also call Claude through https://api.zukko.pro/v1/chat/completions, but then there is no cache discount (more below).
3. Claude Code
Claude Code is Anthropic's command-line coding agent. Install it the official way described in the Claude Code docs (for example, npm install -g @anthropic-ai/claude-code) and set the environment variables:
export ANTHROPIC_BASE_URL=https://api.zukko.pro
export ANTHROPIC_AUTH_TOKEN=fetch_your_key
export ANTHROPIC_MODEL=claude-sonnet-5
export ANTHROPIC_DEFAULT_HAIKU_MODEL=claude-sonnet-5
claudeIt has to be ANTHROPIC_AUTH_TOKEN: Claude Code sends its value in the Authorization: Bearer header, which the gateway accepts. The address has no /v1. ANTHROPIC_DEFAULT_HAIKU_MODEL is required: Claude Code sends background requests to the Haiku-slot model (by default claude-haiku-*), which is not in the Zukko catalog, and the gateway answers an unknown model name with 404 model_not_found. If you switch with /model opus or /model sonnet, also set ANTHROPIC_DEFAULT_OPUS_MODEL=claude-opus-5-5 and ANTHROPIC_DEFAULT_SONNET_MODEL=claude-sonnet-5. To avoid setting the variables in every terminal, you can put them in the env block of ~/.claude/settings.json. Step-by-step setup for macOS, Linux and Windows is in the Claude Code docs.
Claude models and prices
Prices are in USD per 1M tokens, as of October 7, 2026. The intelligence index is the Artificial Analysis Intelligence Index from the Zukko catalog; a dash means no data.
| Model | id | Input / output | Cache read / write | Index | Official price input / output |
|---|---|---|---|---|---|
| Claude Opus 5.5 | claude-opus-5-5 | 0.8 / 4 | 0.08 / 1 | 58 | 4 / 20 |
| Claude Sonnet 5 | claude-sonnet-5 | 0.4 / 2 | 0.05 / 0.5 | — | 2 / 10 |
| Claude Fable 5.1 | claude-fable-5-1 | 7 / 35 | 0.7 / 8.75 | 53 | 10 / 50 |
| Claude Sonnet 4.6 | claude-sonnet-4-6 | 0.5 / 2.5 | 0.063 / 0.6 | — | 3 / 15 |
All of these models have a 1M-token context and accept image input. Provider prices change often: on October 6, 2026, Claude became roughly twice as expensive. For the current list and prices, call GET https://api.zukko.pro/v1/models or open the Models page in the dashboard.
As a reference point: 1,000 messages of 1,500 input and 500 output tokens. Formula: 1.5 × input + 0.5 × output.
- Sonnet 5: 1.5 × 0.4 + 0.5 × 2 = 0.6 + 1 = $1.60;
- Opus 5.5: 1.5 × 0.8 + 0.5 × 4 = 1.2 + 2 = $3.20;
- Fable 5.1: 1.5 × 7 + 0.5 × 35 = 10.5 + 17.5 = $28.
This is a sample calculation without caching; your token volume will differ.
Opus 5.5 or Sonnet 5
Opus 5.5 has the highest intelligence index in the catalog (58) and costs twice as much as Sonnet 5 for both input and output. There is no index data for Sonnet 5. Fable 5.1, with an index of 53, costs far more than either. A sensible order: start with Sonnet 5, run 10–20 of your real tasks on it, and move to Opus 5.5 where the result falls short.
Caching: why the request format matters for Claude
For Claude, prompt caching is billed separately: cache reads cost roughly a tenth of the input price (for Opus 5.5, $0.08 versus $0.8), while cache writes cost more than input ($1). Only the Anthropic format (/v1/messages, the Anthropic SDK, Claude Code) gives the cache discount: there the gateway sees how much input came from the cache. In the OpenAI format, all Claude input is billed at the input price.
For agents like Claude Code this is noticeable. Sample session: 100K fresh input, 2M cache-read tokens, 200K cache-write tokens and 50K output. Formula: (fresh input × input price + cache reads × read price + cache writes × write price + output × output price) / 1,000,000.
- Sonnet 5: 0.1 × 0.4 + 2 × 0.05 + 0.2 × 0.5 + 0.05 × 2 = 0.04 + 0.1 + 0.1 + 0.1 = $0.34;
- Opus 5.5: 0.1 × 0.8 + 2 × 0.08 + 0.2 × 1 + 0.05 × 4 = 0.08 + 0.16 + 0.2 + 0.2 = $0.64.
If the same volume (2.3M input) were billed without the cache, it would come to 2.3 × 0.4 + 0.05 × 2 = $1.02 for Sonnet 5 and 2.3 × 0.8 + 0.05 × 4 = $2.04 for Opus 5.5. The volumes are illustrative: how much lands in the cache depends on the session; a monthly estimate is in the article how much Claude Code costs per month. More in the key limits and cache docs.
Paying from Russia
The balance is kept in dollars, top-ups start at $1:
- SBP and a Russian bank card (Mir, Visa, Mastercard) — in rubles at the market rate at the time of payment;
- cryptocurrency (USDT and others);
- a foreign Mastercard EU card — about a 13% fee, amounts from $2 to $594.
More on paying for Claude specifically in the article how to pay for the Claude API from Russia. To keep an agent from spending more than you planned, set a daily and monthly dollar limit on the key: once it is reached, the gateway returns a 429 error with key_daily_limit_reached or key_monthly_limit_reached.
FAQ
Do I need an Anthropic account?
No. In the chat, via the API and in Claude Code you work with a Zukko key and balance. For Claude Code, ANTHROPIC_BASE_URL, ANTHROPIC_AUTH_TOKEN and catalog models in ANTHROPIC_MODEL and ANTHROPIC_DEFAULT_HAIKU_MODEL are enough.
Why does Claude Code ask me to log in to Anthropic?
Most likely the variables were not picked up: check that they are set in the same terminal where you run claude, that you use ANTHROPIC_AUTH_TOKEN, and that the address has no /v1.
Which Claude model should I start with?
Sonnet 5: it costs half as much as Opus 5.5. If the results on your tasks fall short, switch to claude-opus-5-5.
Can I use Claude for free?
There is no free tier. There is a one-time $0.50 for subscribing to the @zukkopro Telegram channel, once per Telegram account.
zukko
Pay 25×
less
for Claude Code & ChatGPT
One fetch_* key and base URL api.zukko.pro — for Cursor, Claude Code, Codex, and any SDK.
Related articles
- Gemini API billing: how to pay for Gemini by the token from RussiaHow Google bills the Gemini API, where setup gets stuck on payment, and how to call Gemini from a single dollar balance instead.
- xAI Grok API: Connecting Grok 4.7, Price and the 256K ContextHow to call grok-4.7 through an OpenAI-compatible API, what to do about the 256K context limit and what typical requests cost. Plus a comparison with Grok 4.6.
- GPT-6 Astra or Claude Fable 5.1: Two Top Models by Price and IndexTwo premium models with the same intelligence index and different Zukko prices. We price a single request and an agent session with cache, and explain how to choose on your own data.
- Qwen or DeepSeek via API: Price, Context and Intelligence IndexFour models, one task and an honest calculation: what a run costs on Qwen 3.8 Max and Flash versus DeepSeek V4 Pro and V4.1 Flash.