RelayVano API Relay Pricing for Gemini CLI and Codex Users

Field trips EN2026-11-03

One Relay, Several Model Families

Many independent developers start with a familiar assumption: a single API key means a single model, and adding a second model means adding a second provider, a second invoice and a second set of credentials. RelayVano was built to make that assumption less useful. We are an independent relay made for developers worldwide, and one API key reaches every model through https://api.relayvano.com.

That key works across three protocol paths. The OpenAI-compatible /v1 path covers GPT, DeepSeek and the other non-Claude, non-Gemini models. The Anthropic /v1/messages path serves the Claude family. The native Gemini /v1beta path serves Gemini models. We did not flatten these into a single pretend-standard, because the clients developers actually use each expect their own protocol.

The whole site, dashboard and docs come in six languages: 中文, English, Español, हिन्दी, Português, Français. Prices show in yuan on the Chinese pages and in approximate dollars on the others, so a developer reading in any of those languages can check the cost of a plan before committing to it.

Client Pairings Without Guesswork

The second assumption worth challenging is that models are interchangeable once they sit behind a relay. They are not, and pretending otherwise is how people end up debugging a request that was never going to work. Here is the pairing we support, with no exceptions:

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

In practice, setup takes a few decisions. If you are working in Claude Code or Claude Desktop, you point the client at our Anthropic-style endpoint and pick a Claude model. If you are working in Gemini CLI, you set the base address for the native Gemini path and choose a Gemini model. If you are in Codex, Cursor, Cherry Studio, Cline, OpenCode, an OpenAI SDK or plain cURL, you use the OpenAI-compatible path and select GPT, DeepSeek or another model from that group.

RelayVano gives you one account and one balance across all of it, while keeping the protocol each client expects. Claude models belong to Claude Code and Claude Desktop, Gemini models belong to Gemini CLI, and the OpenAI-compatible clients use the GPT, DeepSeek and other non-Claude, non-Gemini models. We would rather spell that out plainly than let you spend an afternoon on a pairing that was never supported.

Pay-as-You-Go Billing You Can Read

The third common assumption is that a relay hides its pricing behind a vendor's subscription, a monthly allowance or a bundle you pay for whether you use it or not. RelayVano sells nothing but pay-as-you-go balance. You pay per token, there is no monthly fee, and the balance never expires.

Prompt tokens, reply tokens, cache stores and cache hits are priced separately, and every request shows up with its charge on the usage page. That last detail matters more than it sounds. When a bill arrives, you can look at the request, see what it cost and understand which part of the traffic drove the number.

Here is a simple example, and it is only an example: say you send 200 requests a day, each with 1,000 prompt tokens and a 500-token reply, using GPT-6.1 Sol. At $0.1958 per million prompt tokens and $0.9792 per million reply tokens, the prompt side works out to about $0.0392 for the day and the reply side to about $0.0979, or roughly $0.137 in total. Change the model and the arithmetic changes with it, which is exactly why per-token pricing suits a solo developer whose traffic is rarely identical from one day to the next.

The fares below are approximate US dollars per million tokens, and the charge itself is made in yuan:

Claude Fable 5.1: $4.07 prompt / $20.33 reply / $0.2222 cache hit, against an official list price of $10 / $50 / $0.25 cache hit.
Claude Opus 5.5: $1.65 prompt / $7.82 reply / $0.075 cache hit, against an official list price of $4 / $20 / $0.2 cache hit.
Claude Sonnet 5.5: $1.25 prompt / $6.25 reply / $0.0889 cache hit, against an official list price of $2 / $10 / $0.1 cache hit.
GPT-5.6 Sol: $0.3736 prompt / $2.09 reply / $0.0306 cache hit, against an official list price of $4 / $20 / $0.4 cache hit. The official promotion is stated to run at least until 2026-11-21.
GPT-6.1 Sol: $0.1958 prompt / $0.9792 reply / $0.0389 cache hit, against an official list price of $2 / $10 / $0.1 cache hit.
Gemini 3.1 Pro: $1.58 prompt / $9.13 reply / $0.1569 cache hit.
Gemini 3.8 Flash: $1.20 prompt / $5.40 reply / $0.1194 cache hit.
DeepSeek V4 Flash: $0.3097 prompt / $1.24 reply / $0.008333 cache hit.
MiniMax M3: $0.1472 prompt / $0.5889 reply / $0.0292 cache hit.

We are an independent service, not an official service, partner or reseller of any model maker. The official list prices above are shown for comparison only.

Interrupts, Top-Ups and Refunds

Another assumption we hear is that a request cut off halfway is a request you do not pay for, or worse, one that runs forever in the background. Neither is how RelayVano behaves. If a caller disconnects mid-stream, the request is still finished and billed on the tokens really used. You get a clean, honest line item rather than a half-generated answer with an ambiguous charge attached.

Adding money is straightforward. Top-ups go through Stripe checkout, where cards are accepted everywhere, plus WeChat Pay where your account supports it. If you change your mind, unspent top-ups can be refunded within 9 days. There is no monthly fee to cancel and no balance that quietly expires while you decide what to build next.

New accounts start with $0.40 of credit, enough to send a few requests and confirm that your client, key and model choice are configured correctly before you top up.

There is also a referral program, because independent developers often know other independent developers. You earn 16% of every top-up by people who sign up through your link, and 4% one level further. Earnings can be withdrawn from $5.56 or moved into your balance and used for requests straight away.

Support From Inside the Dashboard

The last assumption worth breaking is that support disappears once you are small. It does not here. Logged-in members write to us under Messages in the dashboard, and you can keep the whole conversation, the model you were testing and the usage figures in one place.

Shipping alone already means making enough decisions on your own. Picking a relay should not be another source of guesswork. With RelayVano, you get one key, one balance that never expires, transparent per-token charges, six-language documentation and a clear answer about which client speaks to which model. Sign up today, start with the $0.40 of credit and see how your own traffic looks on the usage page.