GPT-6.1 Sol on BestRelay: pay-as-you-go API relay pricing

BestRelay answers with the model you name, not a substitute

Picture an ordinary Tuesday at your own desk. You open BestRelay, create a key, and send a request to GPT-6.1 Sol. BestRelay is an independent API relay, and 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 single property is the reason to keep BestRelay open in the first place: when you ask for a particular model, you get that model back, with its own rate card attached.

The line-up is curated. BestRelay carries 20 public models, each listed under its own name with four prices: input, output, cache write and cache read. When a provider retires a model id, BestRelay delists it rather than leaving a stale name in the catalogue. The rest of this article is a plainly labelled example, following one model, GPT-6.1 Sol, through a sample day so that you can see what the free credit and the rate card look like in ordinary use.

The $0.35 opening credit is enough to make a start on GPT-6.1 Sol

Each new BestRelay account opens with $0.35 of credit. That is the entire allowance, and it is a sensible place to begin: enough to send real requests, read the answers, and inspect the usage records before committing anything further. For example, the free credit alone covers about 1.6 million GPT-6.1 Sol input tokens, which is a useful figure to hold in mind when you decide how hard to test the service.

There is no monthly fee, no subscription and no minimum spend. The balance is prepaid and pay-as-you-go: 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 an untouched balance stays untouched rather than quietly becoming a commitment. Top-ups run through Stripe checkout in the amounts $1.39, $6.94, $13.89, $27.78, $69.44 and $138.89, which lets you size the next step to the work you actually have in front of you.

Two details make the arithmetic easier to trust. Requests the provider rejects are recorded at zero, and when the provider cannot be reached, nothing is charged. Every call also appears in the usage records with its model, tokens and charge, so you can check a figure rather than take it on faith.

A sample desk day with GPT-6.1 Sol, priced as an example

Here is an example, and it is only an example. Take a day in which you send 500,000 input tokens and 100,000 output tokens to GPT-6.1 Sol. On BestRelay, that day costs $0.22, and 22 such days cost $4.74. Now suppose 350,000 of that input is read from cache: the same day costs $0.16.

The published rate for GPT-6.1 Sol is $0.2083 per 1,000,000 input tokens, $1.11 per 1,000,000 output tokens and $0.0431 per 1,000,000 cache read tokens. At the official list price, the same example day comes to $2.00, the 22 days to $44.00 and the cached day to $1.33. The difference on the day is $1.78, and on the 22 days it is $39.26.

Those figures describe the example, not a forecast of your traffic. Your own totals will depend on how many tokens you send, how much of your input is cacheable, and how long your replies run. The point of publishing the rate card is that you can do the same arithmetic yourself before you start, and then confirm it afterwards in the usage records.

BestRelay keeps the client rules and the price list in plain view

BestRelay serves Claude models only to the official Claude Code and Claude Desktop clients, version 2.1.285 or newer. Those models 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. Stating this plainly is kinder than discovering it later, because a client that speaks the wrong protocol will simply never work with the model you had in mind.

Connecting a client takes two settings: the address and a key created under API keys in the dashboard. The details live in the BestRelay docs. An account can hold several keys, each one can be voided at any time, and each key is limited to 120 requests a minute, with 8 calls in flight on one account. Keeping a separate key for a separate project makes the usage records easier to read at the end of the month.

On privacy, BestRelay keeps the user-written text of requests for 30 days, so abuse can be reviewed and billing questions settled. System prompts and model replies are not stored. The request text is the part you wrote; the answer is left out of the record.

Top-ups, refunds and referrals, stated without fine print

For example, if an invited member tops up $13.89, the two commissions are $2.50 and $0.5556: 18% of the top-ups of people who joined through your link, and 4% of the top-ups of the people they invited. Commission can be moved into your BestRelay balance, or withdrawn to Alipay from $20.83. A referral arrangement is only worth anything if the numbers are legible, so here they are in plain figures rather than in a paragraph of percentages.

Unused credit can be refunded on request within 13 days of the top-up, less what was consumed and any payment-processor fees. Gift credit and commission cannot be refunded, which is worth keeping in mind when you move commission into the balance and then treat that balance as ordinary spendable credit.

What BestRelay adds up to is a plain service: a curated line-up, published rates, a small opening credit and a request that answers under the name you gave it. If you would like to see how the day feels at your own desk, open an account and put the first request through GPT-6.1 Sol.