Claude Code and BestRelay: a two-setting API relay setup

BestRelay is an independent relay where the model you name is the model that answers

BestRelay is an independent API relay, and the promise it makes is plain: the model you name is the model that answers. Each request is forwarded to the model provider under the model name the caller wrote, so nothing is quietly swapped for a cheaper tier behind the scenes. That distinction matters in a field where the label on a request and the model that produces the reply can drift apart. On BestRelay they do not.

The line-up is curated rather than accumulated: 20 public models, each listed under its own name, each carrying four prices, namely input, output, cache write and cache read. When a provider retires a model id, BestRelay delists it, so the card does not keep advertising something that has already gone away. Keeping the selection short is what allows the card to be read line by line rather than skimmed.

Money is prepaid. You hold a balance, you spend from it, and there is no monthly fee, no subscription and no minimum spend. Nothing is held in advance against a call you have not made yet. A new account opens with $0.35 of credit, enough to send a request and read the resulting usage record before you decide to add anything to the balance.

BestRelay also keeps a paper trail. Every call appears in the usage records with its model, tokens and charge. An account can hold several keys, and any of them can be voided at any time. The relay stands apart from the providers it forwards to, and it does not rename what you asked for.

The rate card reads cleanly because every model carries four prices

All prices are per 1,000,000 tokens, and each line names its model before it names its numbers. The Claude entries read like this:

• Claude Fable 5: input $4.17, output $20.97, cache write $5.14, cache read $0.3472. The official list price is input $10, output $50, cache read $1.
• Claude Fable 5.1: input $4.10, output $20.42, cache write $5.00, cache read $0.2083. The official list price is input $10, output $50, cache read $0.25.
• Claude Haiku 4.5: input $0.4167, output $2.01, cache write $0.4861, cache read $0.0333. The official list price is input $1, output $5, cache read $0.1.
• Claude Opus 5: input $2.29, output $11.46, cache write $2.85, cache read $0.1389. The official list price is input $5, output $25, cache read $0.5.
• Claude Opus 5.5: input $1.74, output $8.13, cache write $2.01, cache read $0.0778. The official list price is input $4, output $20, cache read $0.2.
• Claude Sonnet 5: input $1.32, output $6.67, cache write $1.67, cache read $0.1167. The official list price is input $2, output $10, cache read $0.2.
• Claude Sonnet 5.5: input $1.39, output $7.01, cache write $1.74, cache read $0.0944. The official list price is input $2, output $10, cache read $0.1.

The rest of the card follows the same convention, with cache write listed where the provider reports one: DeepSeek V4 Pro at input $1.53, output $4.58, cache read $0.0514; Gemini 2.5 Flash at input $0.2778, output $2.08, cache write $0.2778, cache read $0.0264; Gemini 3.1 Pro at input $1.74, output $9.93, cache write $1.74, cache read $0.1389; Gemini 3.8 Flash at input $1.25, output $5.56, cache write $1.25, cache read $0.1236; Gemini 4 Argon at input $1.67, output $8.06, cache write $1.67, cache read $0.0847; Kimi K3 at input $1.39, output $7.01, cache read $0.1389; GPT-5.6 Sol at input $0.4167, output $2.29, cache read $0.0333; GPT-6 Astra at input $0.7639, output $3.96, cache read $0.0792; GPT-6.1 Sol at input $0.2083, output $1.11, cache read $0.0431; Qwen3.8 Max at input $0.8333, output $2.57, cache read $0.1014; Grok 4.6 at input $0.625, output $1.81, cache read $0.1389; Grok 4.7 at input $0.625, output $1.94, cache read $0.1389; GLM-5.3 at input $0.625, output $1.94, cache read $0.1389.

Claude Code needs two settings and a key, and the client rules are explicit

Connecting a client to BestRelay takes two settings: the address and a key created under API keys in the dashboard. The details live in the documentation, and the address itself is deliberately absent from this article, along with every other setting value. Nothing about the setup is secret in the sense of being hidden from you; it is simply kept out of prose that gets copied and pasted around.

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, and BestRelay states this rather than leaving you to find out by trial and error. Gemini models use their native protocol, through Gemini CLI or direct generateContent calls, and are not OpenAI-compatible. Tools such as Codex pair with GPT, DeepSeek, Kimi, GLM, Qwen and Grok models instead.

Two limits sit alongside the setup. A single key carries 120 requests a minute, and an account may have 8 calls in flight at once. An account can hold several keys, each of which can be voided at any time, so a key that travels further than you intended can be retired without disturbing the rest of your work. Read the card, set the two settings, and the relay does the rest of the forwarding under the model name you typed.

Billing is pay-as-you-go, charged once after each call ends

Billing on BestRelay is pay-as-you-go, and the sequence is short. You top up, the balance sits there, 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 an idle balance is not a commitment, and there is no monthly fee to think about.

Top-ups run through Stripe checkout in fixed amounts: $1.39, $6.94, $13.89, $27.78, $69.44 and $138.89. The smaller amounts suit a trial run; the larger ones suit a working routine. Refunds are available on unused credit 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 referral earnings are treated as ordinary balance.

Failure has its own rule. A request the provider rejects is recorded at zero, and when the provider cannot be reached at all, nothing is charged. You are not billed for a reply that did not arrive, and the usage records show it rather than quietly absorbing the cost.

For example, a day of 500,000 input tokens and 100,000 output tokens on Claude Sonnet 5.5 costs $1.40 on BestRelay; 22 such days cost $30.71, and 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 is $0.60 on the day and $13.29 across the 22 days.

Records, privacy and referrals keep the ledger legible

Every call on BestRelay appears in the usage records with its model, tokens and charge, which means you can check a bill line against the work that produced it. Privacy is handled in a deliberately narrow way: the user-written text of requests is kept for 30 days to review abuse and settle billing questions, while system prompts and model replies are not stored.

Referrals are paid on the top-ups of people who join through your link. 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. For example, on an invited member's top-up of $13.89, the two commissions are $2.50 and $0.5556.

The arithmetic at the low end of the card is easy to check. For example, a day of 500,000 input tokens and 100,000 output tokens on GPT-6.1 Sol 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 is $1.78 on the day and $39.26 across the 22 days. The $0.35 of free credit on a new account covers about 1.6 million GPT-6.1 Sol input tokens, which is a quiet way to see the rate card in motion before you commit to a larger top-up. When you are ready, BestRelay is ready with the same card, the same keys and the same records.