BestRelay and Claude Code: what pay-as-you-go costs

BestRelay has no monthly fee, and the model you name is the model that answers

Here is a sample day at your own desk, offered as an example rather than a promise about how your own day will run. You open BestRelay in the morning, pick a model from the curated line-up of 20 public models, and start writing. When you ask for Claude Opus 5.5, BestRelay forwards the request to the provider under that model name. The model named is the model that answers, and nothing is quietly swapped for a cheaper tier.

There is no subscription and no minimum spend. You keep 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. If a request is rejected by the provider, it is recorded at zero, and when the provider cannot be reached, nothing is charged.

A new account opens with $0.35 of credit, which is enough to make a start before you top up. For example, that free credit alone covers about 1.6 million GPT-6.1 Sol input tokens, so you can see the shape of the service before spending a cent of your own money.

The point of the sample day is quiet. You are not managing a plan, watching a quota reset, or asking a sales team whether a feature is included. You are choosing a model, sending text, and reading a charge that matches the tokens the provider reported. BestRelay keeps the records: every call appears in the usage records with its model, tokens and charge.

BestRelay lists what a working day costs, with prices shown per model

Every model in the BestRelay line-up is listed under its own name with four prices: input, output, cache write and cache read, given per 1,000,000 tokens. Take Claude Sonnet 5.5 as an example. A day of 500,000 input 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, 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; the difference on the day is $0.60 and on the 22 days $13.29.

GPT-6.1 Sol makes the same point in a different currency. As an example, a day of 500,000 input 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 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.

The smaller entries earn their place in the list too. Claude Haiku 4.5 is $0.4167 input, $2.01 output, $0.4861 cache write and $0.0333 cache read per 1,000,000 tokens. GPT-5.6 Sol is $0.4167 input, $2.29 output and $0.0333 cache read. Gemini 2.5 Flash is $0.2778 input, $2.08 output, $0.2778 cache write and $0.0264 cache read.

None of this needs a plan tier to be negotiated. You choose the model, and the rate card is what you pay.

The BestRelay line-up holds 20 public models, each named and priced on its own

The line-up is curated: 20 public models, each listed under its own name with its own four prices. The Claude family is there — Claude Fable 5, Claude Fable 5.1, Claude Haiku 4.5, Claude Opus 5, Claude Opus 5.5, Claude Sonnet 5 and Claude Sonnet 5.5. The GPT family includes GPT-5.6 Sol, GPT-6 Astra and GPT-6.1 Sol. The Gemini models include Gemini 2.5 Flash, Gemini 3.1 Pro, Gemini 3.8 Flash and Gemini 4 Argon. There is also DeepSeek V4 Pro, Kimi K3, Qwen3.8 Max, Grok 4.6, Grok 4.7 and GLM-5.3.

Prices differ across that line-up because the models differ. Claude Opus 5.5 is $1.74 input, $8.13 output, $2.01 cache write and $0.0778 cache read. Claude Fable 5 is $4.17 input, $20.97 output, $5.14 cache write and $0.3472 cache read. Gemini 3.1 Pro is $1.74 input, $9.93 output, $1.74 cache write and $0.1389 cache read. Kimi K3 is $1.39 input, $7.01 output and $0.1389 cache read. Grok 4.7 is $0.625 input, $1.94 output and $0.1389 cache read.

Curated also means maintained. When a provider retires a model id, BestRelay delists it, so the list does not keep offering a name that no longer answers. What remains on the line-up is what can be called today.

For a reader planning an afternoon of work, this is a useful kind of certainty: the name on the card is the name in the request, and the price beside it is the price that will be charged.

BestRelay fits the client you already use, and the setup is two settings

Claude models are served 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. Tools like Codex pair with GPT, DeepSeek, Kimi, GLM, Qwen and Grok models. Gemini models use their native protocol, through Gemini CLI or direct generateContent calls, and are not OpenAI-compatible.

Connecting a client takes two settings: an address and a key created under API keys in the dashboard. The details are in the BestRelay docs. Beyond that, the limits are plain: 120 requests a minute on one key, and 8 calls in flight on one account. Your own desk in the sample day is not a special case; those are the numbers the service works to.

An account can hold several keys, and each can be voided at any time, so a key you no longer trust can be closed without disturbing the rest of your setup. Every call appears in the usage records with its model, tokens and charge, which means a charge you did not expect can be traced back to the model that produced it.

On privacy, BestRelay keeps the user-written text of requests for 30 days, to review abuse and settle billing questions. System prompts and model replies are not stored. The practical effect for you is that your usage records show what you asked and what it cost, without the service keeping a second copy of the conversation.

Top-ups, refunds and referrals, all handled from the same prepaid balance

Top-ups run through Stripe checkout in these amounts: $1.39, $6.94, $13.89, $27.78, $69.44 and $138.89. For example, a $13.89 top-up is the point at which the referral arithmetic becomes easy to follow: on an invited member's top-up of $13.89, the two commissions are $2.50 and $0.5556. You receive 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 the balance, or withdrawn to Alipay from $20.83.

Refunds are available on unused credit 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.

Back to the sample day. You started with $0.35 of credit, worked through the afternoon, and by the evening the usage records show each call with its model, tokens and charge. If a request was rejected, it was recorded at zero; if the provider could not be reached, nothing was charged. You topped up in the amount that suited you, and the balance moved only as calls ended.

That is the whole arrangement, and it is deliberately plain: no monthly fee, no subscription, no minimum spend, and a line-up where the model you name is the model that answers. If you would like to see how that feels on your own desk, BestRelay is ready when you are.