RelayVano API Relay Pricing: GPT-6.1 Sol with $0.40 Starter Credit

Match your client to the model family before you sign up

Before you create a RelayVano account, the first box to tick on your decision checklist is simple: which coding client do you already use, and which model family do you want to learn with? This matters because one API key on RelayVano reaches every model through https://api.relayvano.com, but each model family speaks its own protocol. A key that works in one tool does not automatically work in another, and guessing at this stage is where beginners usually lose an afternoon.

Claude models work only in the official Claude Code and Claude Desktop clients, version 2.1.285 or newer. If your plan is to study with Claude, make sure one of those two clients is installed and up to date before you set anything up. Gemini models use the native Gemini protocol only, in Gemini CLI, and they are not OpenAI-compatible. Codex, Cursor, Cherry Studio, Cline, OpenCode, OpenAI SDKs and cURL pair only with GPT, DeepSeek and the other non-Claude, non-Gemini models.

That pairing is not a suggestion or a preference; it is the way the models are built. So the practical checklist item becomes: choose a client you like, then choose a model family that belongs in it. Once the pair is right, setup reduces to pointing the client at our base address and adding your key. RelayVano gives you a single key to work with, which keeps your student project tidy instead of scattering credentials across a dozen tools.

Compare pay-as-you-go fares with a vendor subscription

RelayVano sells prepaid balance and nothing else. You pay per token, there is no monthly fee, and the balance never expires. For a student who codes in bursts rather than every single day, that structure is often easier to reason about than a vendor's subscription: you are not paying for idle weeks, and leftover credit is still there when the next assignment starts.

The per-million-token fares give you a concrete comparison. Take Claude Sonnet 5.5 at $1.25 prompt, $6.25 reply and $0.0889 cache hit, where the official list price is $2 / $10 / $0.1 cache hit. GPT-6.1 Sol is listed at $0.1958 prompt, $0.9792 reply and $0.0389 cache hit, against an official list price of $2 / $10 / $0.1 cache hit. For lighter experiments, MiniMax M3 is $0.1472 prompt, $0.5889 reply and $0.0292 cache hit, and MiMo V2.6 Pro is $0.2319 prompt, $0.4653 reply and $0.002778 cache hit. These are approximate US dollar figures, with the charge itself made in yuan.

One detail worth reading carefully: prompt tokens, reply tokens, cache stores and cache hits are priced separately. That means long context files and repeated reference material can behave differently from a short question, and it means the usage page can explain exactly where a charge came from. When you compare RelayVano with a vendor's subscription, compare it with the full picture of your own usage, not with a headline number.

Plan a student budget with the $0.40 starter credit

New accounts start with $0.40 of credit. That is small on purpose, and it is genuinely useful: it gives you enough room to send a handful of real requests, read the responses, and confirm that your client, your key and your chosen model all work together before you commit any money.

Here is an example to help you size things. Suppose you send 200 requests a day, each with 500 prompt tokens and 500 reply tokens, using GPT-6.1 Sol and ignoring caching for simplicity. At $0.1958 per million prompt tokens and $0.9792 per million reply tokens, one request costs about $0.0005875. That works out to roughly $0.1175 per day and about $3.525 over 30 days. This is an example calculation, not a prediction of what you will spend; your own prompts and answers will be different lengths, and caching can shift the mix.

Two habits keep that number predictable. First, check the usage page, where every request appears with its charge. Second, remember that if a caller disconnects mid-stream, the request is still finished and billed on the tokens really used. That is why a runaway loop is worth stopping early rather than letting it run while you make tea. Small top-ups, sized from what the usage page actually shows, are a comfortable way to learn: you top up, you watch, you adjust.

Look closely at top-ups, refunds and balance rules

Top-ups go through Stripe checkout. Cards work everywhere, and WeChat Pay is available where your account supports it, which is handy for students whose budget sits in different payment rails. Unspent top-ups can be refunded within 9 days, so a mistaken large top-up is not a permanent commitment.

The balance itself does not expire. There is no monthly fee to forget about, and no recurring charge quietly accumulating in the background while you experiment. You add credit when you need it, and the amount you have left stays available for the next session. For a learner moving through a course, a semester project and then a personal build, that continuity is a real advantage over starting from zero each month.

Our site, dashboard and docs come in six languages: 中文, English, Español, हिन्दी, Português and Français. Prices show in yuan on the Chinese pages and in approximate dollars on the others, so you can read the numbers in the language and currency convention that suits you. RelayVano is an independent relay made for developers worldwide, and the multilingual setup is part of that promise rather than a decorative extra.

Know where to go when something does not work

Checklist items should include the boring stuff, because that is where friction hides. If a request behaves oddly, open the usage page and check whether the call was recorded and what it cost. If the pairing is wrong, re-read the client rules above: Claude belongs in Claude Code or Claude Desktop, Gemini belongs in Gemini CLI, and the other clients take the non-Claude, non-Gemini models.

If you still need a hand, logged-in members write to us under Messages in the dashboard. That keeps support attached to the same account as your balance and your usage history, so you do not have to reconstruct your setup from memory in a support form. It also means the conversation stays available while you iterate.

Run through this checklist, then start with $0.40

Here is the short version you can work through before you decide:

• Confirm the client and model pairing: Claude Code or Claude Desktop for Claude, Gemini CLI for Gemini, and the other listed clients for GPT, DeepSeek and the remaining non-Claude, non-Gemini models.
• Check that the client version is recent enough, including version 2.1.285 or newer for Claude Code and Claude Desktop.
• Compare the pay-as-you-go fares for the models you actually plan to use, including prompt, reply and cache pricing, against what a vendor's subscription would cost you.
• Start with the $0.40 credit, send a few real requests, and read the charges on the usage page before topping up.
• Size top-ups from observed usage, keep the option of a refund within 9 days, and remember the balance does not expire.
• Check that the six-language site, dashboard and docs work for you, and note that Messages in the dashboard is where support lives.
• If you want to share RelayVano with classmates, note that people who sign up through your link contribute 16% of every top-up to you, 4% one level further, and earnings can be withdrawn from $5.56 or moved into your balance.

Learning to code is expensive enough in time. Paying for it should not have to be complicated. With one API key, transparent per-token pricing, small top-ups you control and a $0.40 starter credit to prove the whole thing works, RelayVano keeps the decision low-risk. Read the checklist, pick the client and model pair that fits your coursework, and sign up today.