Log in·Sign up
BestRelay

Back to the Journal

BestRelay setup for GPT-6.1 Sol on a pay-as-you-go API relay

BestRelay answers with the model you name, not a substitute

Here is an example day at your own desk, and how BestRelay fits into it. You sit down with a short list of work and a client already open on the screen. BestRelay is an independent API relay: it forwards each request to the model provider under the model name you wrote, so the model named is the model that answers, and nothing is quietly swapped for a cheaper tier. That sentence is the whole commercial argument, and it is worth reading twice.

The line-up is curated rather than sprawling: 20 public models, each listed under its own name with four prices, covering input, output, cache write and cache read. When a provider retires a model id, BestRelay delists it, so a stale name does not sit in the list waiting to mislead you. The naming discipline carries into the clients. Claude models are served only to the official Claude Code and Claude Desktop clients, version 2.1.285 or newer, and they do not work in Cursor, Cherry Studio, Cline, OpenCode, any SDK or cURL. Gemini models use only their native protocol, through Gemini CLI or direct generateContent calls, and are not OpenAI-compatible.

Take pricing as an example. On Claude Sonnet 5.5, a day of 500,000 input and 100,000 output tokens costs $1.40 on BestRelay; 22 such days cost $30.71; with 350,000 of the input read from cache the day costs $0.94. At the official list price the same day is $2.00, the 22 days $44.00 and the cached day $1.33, so the difference on the day is $0.60 and on the 22 days $13.29.

BestRelay bills a prepaid balance once per finished call

On BestRelay the account runs on a prepaid pay-as-you-go balance. There is no monthly fee, no subscription and no minimum spend, and each call is charged once, after it ends, from the tokens the provider reports, at the rate card prices. Nothing is held in advance, so a quiet morning does not quietly become a bill.

The example day runs cleanly because of that arrangement. A new account opens with $0.35 of credit, and the free credit alone covers about 1.6 million GPT-6.1 Sol input tokens. As an example, a day of 500,000 input and 100,000 output tokens costs $0.22 on BestRelay, 22 such days cost $4.74, and with 350,000 of the input read from cache the day costs $0.16. At the official list price the same day is $2.00, the 22 days $44.00 and the cached day $1.33, so the difference on the day is $1.78 and on the 22 days $39.26.

Failures are handled without drama. Requests the provider rejects are recorded at zero, and when the provider cannot be reached nothing is charged. Top-ups go through Stripe checkout, and the available amounts are $1.39, $6.94, $13.89, $27.78, $69.44 and $138.89. You pick the size that suits the pace of the example day, and the balance carries the rest. Every call appears in the usage records with its model, tokens and charge, so at the end of the day you can read what happened rather than guess at it.

BestRelay referral commission is 18% and 4% of invited top-ups

Referrals are the part of the example day that starts to look like a small income stream. The arrangement is plain: 18% of the top-ups of people who joined through your link, and 4% of the top-ups of the people they invited. As an example, on an invited member's top-up of $13.89, the two commissions are $2.50 and $0.5556.

The example day also shows where the boundaries sit. Gift credit and commission cannot be refunded, so treat them as yours to spend or withdraw rather than as money you can hand back. Unused credit can be refunded on request within 13 days of the top-up, less what was consumed and any payment-processor fees.

Top-up amounts are the same list you saw earlier: $1.39, $6.94, $13.89, $27.78, $69.44 and $138.89, paid through Stripe checkout. If an invited colleague tops up at $13.89, you can work out the 18% and the 4% from the published figures rather than from a promise, and you can watch the same amounts arrive in your account as the day goes on. Nothing about BestRelay requires you to guess how the referral side is calculated.

BestRelay moves commission into the balance or withdraws it to Alipay

Commission can be moved into the balance, or withdrawn to Alipay from $20.83. On the example day, that choice is the practical one: moving the money in means it is available for the next call, while a withdrawal settles it out through Alipay once the balance reaches the threshold. BestRelay does not ask you to wait for a payout cycle or a review that it has not promised.

The rest of the desk is deliberately plain. An account can hold several keys, and each key can be voided at any time, so a key handed to a colleague for an afternoon can be closed when it is no longer wanted. Every call appears in the usage records with its model, tokens and charge, which means the day's spending can be read line by line instead of estimated from a feeling.

Privacy is stated in the same register. The user-written text of requests is kept for 30 days to review abuse and settle billing questions; system prompts and model replies are not stored. That is the whole retention policy described, and it is worth knowing before you paste anything sensitive into a prompt on the example day.

BestRelay setup takes two settings, and the client rules are fixed

Connecting a client takes two settings: the address and a key created under API keys in the dashboard, with the details kept in the docs. The limits are modest and stated plainly: 120 requests a minute on one key, and 8 calls in flight on one account. On the example day those numbers are enough to keep the work moving without turning into a queue.

The pairing rules matter more than the limits. Claude models are served only to the official Claude Code and Claude Desktop clients, version 2.1.285 or newer, and they do not work in Cursor, Cherry Studio, Cline, OpenCode, any SDK or cURL. Tools such as Codex pair only with GPT, DeepSeek, Kimi, GLM, Qwen and Grok. Gemini models use only their native protocol, through Gemini CLI or direct generateContent calls, and are not OpenAI-compatible.

The example day ends with the same point it began with: the model you name is the model that answers. BestRelay keeps the line-up curated at 20 public models, prices it clearly, and lets the day's costs be checked in the usage records. If you would like to see how that behaves at your own desk, an account with BestRelay is a reasonable place to start.