GPT-6.1 Sol on BestRelay: API relay setup with $0.2083 input

BestRelay has no monthly fee, and your balance is the whole bill

Dear reader, welcome to BestRelay, and welcome to a letter rather than a sales page. You asked what a quiet week costs here, so let us answer plainly. 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. There is no monthly fee, no subscription and no minimum spend. You hold a prepaid pay-as-you-go balance, 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.

A quiet week, then, is a week of ordinary work: a few dozen calls, each metered on its own terms. If a provider rejects a request, that request is recorded at zero. If the provider cannot be reached, nothing is charged. You do not pay for a service you did not receive. Behind those calls sits a curated line-up of 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, so the shelf never accumulates names that no longer answer.

Everything else is deliberately plain. An account can hold several keys, and each key can be voided at any time. Every call appears in the usage records with its model, its tokens and its charge. That is the whole arrangement for a quiet week: a balance, a key, a list of models, and a ledger you can read afterwards.

BestRelay forwards the model name you write, unchanged

Because BestRelay is a relay and not a broker, the name in your request is the name on your bill. Ask for Claude Fable 5 and you are talking to Claude Fable 5, at $4.17 per 1,000,000 input tokens, $20.97 for output, $5.14 for a cache write and $0.3472 for a cache read. Ask for Claude Fable 5.1 and the rates are $4.10, $20.42, $5.00 and $0.3472. The point is not that one name is wiser than another; it is that the label you chose is the label you pay for, and the numbers beside it are printed openly.

The same care runs through the rest of the shelf. Claude Haiku 4.5 lists input at $0.4167 and output at $2.01. Claude Opus 5 lists $2.29 and $11.46, while Claude Opus 5.5 lists $1.74 and $8.13. GPT-6.1 Sol lists input at $0.2083, output at $1.11 and cache read at $0.0431; GPT-6 Astra lists $0.7639, $3.96 and $0.0792. Gemini 3.1 Pro lists $1.74 for input and $9.93 for output, Gemini 3.8 Flash $1.25 and $5.56, and Gemini 4 Argon $1.67 and $8.06.

Elsewhere, DeepSeek V4 Pro lists $1.53 and $4.58 with cache read at $0.0514; Kimi K3 lists $1.39 and $7.01; Qwen3.8 Max lists $0.8333 and $2.57; Grok 4.6 lists $0.625 and $1.81; Grok 4.7 lists $0.625 and $1.94; GLM-5.3 lists $0.625 and $1.94. Twenty names, each with its own row, each delisted when the provider retires it. You choose; BestRelay keeps the label honest.

Worked examples: what a week of ordinary calls comes to

Let us put two of those rows to work. Take Claude Sonnet 5.5 as an example: a day of 500,000 input tokens and 100,000 output tokens costs $1.40 on BestRelay, and 22 such days cost $30.71. With 350,000 of the input read from cache, that day costs $0.95. At the official list price the same day is $3.00, the 22 days $66.00 and the cached day $2.06, so the difference is $1.60 on the day and $35.29 across the 22 days.

Now take GPT-6.1 Sol as an example: a day of 500,000 input tokens and 100,000 output tokens costs $0.22 on BestRelay, and 22 such days cost $4.74. With 350,000 of the input read from cache, the day costs $0.16. At the official list price, that day is $2.00, the 22 days $44.00 and the cached day $1.48, a difference of $1.78 on the day and $39.26 across the 22 days.

These are examples, not promises, but they show the shape of a quiet week: a small number of cents or dollars per day, added up as calls are made, with no standing charge waiting at the end of the month. The $0.35 of credit that opens each new account covers about 1.6 million GPT-6.1 Sol input tokens, which is enough to see how the meter behaves before you top up.

Top-ups, refunds and referrals at BestRelay

Adding funds is done through Stripe checkout, in set amounts: $1.39, $6.94, $13.89, $27.78, $69.44 and $138.89. Pick the size that suits a week rather than a quarter, since BestRelay asks for no minimum spend and holds nothing back. If a top-up turns out to be larger than your plans, 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 knowing before you treat a referral balance as spending money.

Referrals are simple to describe. You earn 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 balance, or withdrawn to Alipay from $20.83. To make the arithmetic concrete, on an invited member's top-up of $13.89, the two commissions are $2.50 and $0.5556. The same discipline applies here as everywhere else on BestRelay: the rule is printed, the amounts are fixed, and the ledger shows what happened.

Between them, the free opening credit, the top-up amounts and the referral percentages mean that a new reader can begin small, watch a week of calls accumulate, and decide afterwards how much to hold in the balance.

Clients, keys and records: how BestRelay fits your desk

Some care is needed about which client speaks to which model, and we prefer to say so plainly. Claude models are served only to the official Claude Code and Claude Desktop clients, version 2.1.285 or newer; they do not work in Cursor, Cherry Studio, Cline, OpenCode, any SDK or cURL. Those tools, like 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. Getting this wrong costs you a call, so BestRelay documents it rather than leaving you to guess.

Connecting takes two settings: the address and a key created under API keys in the dashboard. The details are in the docs; we will not print them here, partly because they rarely belong in a letter and partly because yours should stay yours. Limits are modest and stated: 120 requests a minute on one key, and 8 calls in flight on one account.

On privacy, we are brief because the facts are brief. 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. Every call appears in the usage records with its model, tokens and charge, and any key can be voided at any time. That is the whole of BestRelay: no monthly fee, a curated line-up, honest names, and a bill you can read. If a quiet week sounds like the right way to begin, open an account and see how the meter behaves.