RelayVano API Relay Setup with the OpenAI SDK: GPT-6.1 Sol

Coding-agent routes EN2026-11-03

One Address Change Connects the OpenAI Python SDK to RelayVano

If you already write Python against the official OpenAI Python SDK, you do not need to learn a new client to try RelayVano. The switch is a base-address change. Point the SDK's base URL at https://api.relayvano.com, add your RelayVano API key in the place where you normally add your OpenAI key, and the same request objects, streaming calls and response handling keep working as before. There is no SDK rewrite to plan, no wrapper library to install and no migration project to schedule.

Here is the setup in words. First, sign in to RelayVano and create a key from the dashboard. Next, set the SDK's base URL to https://api.relayvano.com, either in code or through the environment variable your project already uses for the endpoint. Then pass the RelayVano key as the API key argument. After that, run your existing script. Requests that previously went to the default OpenAI endpoint now travel through RelayVano, and your usage appears on the usage page with the charge for each request.

RelayVano is an independent relay made for developers worldwide, and the whole site, dashboard and documentation are available in six languages: 中文, English, Español, हिन्दी, Português, Français. Prices are shown in yuan on the Chinese pages and in approximate dollars on the others, which makes the documentation and the dashboard comfortable to read whether you are in Shanghai, São Paulo, Lyon or Delhi. For a Python team, that combination of a familiar SDK and a multilingual control panel is a short path from curiosity to a working call.

Which Models the OpenAI SDK Can Reach Through RelayVano

One API key reaches every model through https://api.relayvano.com, and the same key opens three protocol paths: the OpenAI-compatible /v1, Anthropic /v1/messages and native Gemini /v1beta. The important detail is which client pairs with which models, and the pairing is specific rather than approximate.

Codex, Cursor, Cherry Studio, Cline, OpenCode, OpenAI SDKs and cURL pair with GPT, DeepSeek and the other non-Claude, non-Gemini models.
Claude models work only in the official Claude Code and Claude Desktop clients, version 2.1.285 or newer.
Gemini models use the native Gemini protocol only, through Gemini CLI, and they are not OpenAI-compatible.

In practice, if your application is written against the OpenAI Python SDK, you are in the first group. You can call GPT-6.1 Sol, GPT-6 Astra, GPT-5.6 Sol, DeepSeek V4 Pro, DeepSeek V4 Flash, Qwen3.7 Plus, Qwen3.8 Max, Kimi K3, MiniMax M3, MiMo V2.6 Pro, Hunyuan 4, Grok 4.6 and Grok 4.7 through that same client and the same base address. Claude and Gemini requests follow their own rules, so keep them in their matching clients rather than forcing them into an SDK that does not speak their protocol.

A Worked Example: Every Sum Shown

To make the pricing concrete, here is a worked example. The figures below are an example only, calculated from the published RelayVano fares, and they are not a quote for your own workload. We use GPT-6.1 Sol, whose approximate US dollar fare per million tokens is $0.1958 for prompt tokens, $0.9792 for reply tokens and $0.0389 for cache hits.

Example request: 1,000 prompt tokens and 500 reply tokens in a single call.

Prompt tokens: 1,000 ÷ 1,000,000 = 0.001 million tokens. 0.001 × $0.1958 = $0.0001958.
Reply tokens: 500 ÷ 1,000,000 = 0.0005 million tokens. 0.0005 × $0.9792 = $0.0004896.
Total for the request: $0.0001958 + $0.0004896 = $0.0006854.

Now extend the example. If you send 200 requests a day at that size, that is an example workload, not a typical one: 200 × $0.0006854 = $0.13708 per day. Over 30 days the example totals 30 × $0.13708 = $4.1124. A second example with a different model shows how the same arithmetic moves. GPT-6 Astra is listed at $0.7403 per million prompt tokens and $3.70 per million reply tokens, so a 2,000-token prompt with a 1,000-token reply costs 0.002 × $0.7403 = $0.0014806 for the prompt, plus 0.001 × $3.70 = $0.00370 for the reply, giving $0.0051806 per request. One hundred such requests would be 100 × $0.0051806 = $0.51806.

Cache tokens are priced separately from prompt and reply tokens, so you can include them in the same style of calculation. In the first example, 10,000 cache-hit tokens would be 0.01 × $0.0389 = $0.000389. Because prompt tokens, reply tokens, cache stores and cache hits are each priced on their own line, the usage page shows you exactly which category produced each charge.

What the Pay-as-you-go Balance Changes About Spending

RelayVano sells pay-as-you-go balance and nothing else. You top up a prepaid balance, you pay per token, there is no monthly fee, and the balance never expires. That shape suits workloads whose volume changes week to week: a hobby project that idles for a month does not lose its credit, and a batch job that runs heavily for a day can draw down the balance you actually funded.

Top-ups run through Stripe checkout. Cards work everywhere, and WeChat Pay is available where your account supports it. If you change your mind, unspent top-ups can be refunded within 9 days. New accounts start with $0.40 of credit, which is enough to make a first call and see the request, the reply and the charge on the usage page before you commit to a larger top-up.

One more practical point about billing: if a caller disconnects mid-stream, the request is still finished and billed on the tokens really used. That means a client that closes a connection does not necessarily cancel the work the provider already did, so it is worth treating the usage page as the record of truth when you reconcile a month's spend.

Six Languages, Worldwide Developers, and a Clear Next Step

RelayVano is built for developers worldwide, and the six-language site, dashboard and docs are part of that design rather than a decorative extra. When an error message arrives at three in the morning, reading it in your own language saves a search. When you compare fares, the Chinese pages show prices in yuan while the other pages show approximate dollars, so a team can budget in the currency it already uses and still compare like for like.

You can also share in the growth. Referrals pay 16% of every top-up made by people who sign up through your link, plus 4% one level further. Earnings can be withdrawn from $5.56 or moved straight into your balance, so a referral month can become working credit without a separate cash-out step. If anything about your account needs attention, logged-in members write to us under Messages in the dashboard.

Here is the short version of the whole plan. Sign up for RelayVano, copy your API key, change one address in your OpenAI Python SDK configuration to https://api.relayvano.com, and run the code you already have. Watch the first request and its charge on the usage page, top up with a card or WeChat Pay when it suits you, and add your $0.40 of starting credit to the balance as you go. The SDK stays the SDK, the models stay familiar and the bill follows the tokens you actually use. Create your RelayVano account today and make the one-address change this week.