BestRelay API relay: GPT-6.1 Sol at $0.2083 per million input tokens

Protocols, closely readEN2026-11-03

BestRelay curates a line-up where the model you name is the model that answers

You open BestRelay at your desk on a Tuesday morning. The dashboard shows a curated list of twenty public models, each listed under its own name with four prices: input, output, cache write, and cache read. 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. When a provider retires a model id, BestRelay delists it. You see Claude Fable 5 at $4.17 input and $20.97 output per million tokens, and you see GPT-6.1 Sol at $0.2083 input and $1.11 output per million tokens. The list is finite, deliberate, and priced per million tokens with cache rates where the provider offers them.

You pick a model by its exact name. BestRelay does not rename or alias them. If you send a request for Claude Sonnet 5.5, the call goes to the provider as Claude Sonnet 5.5. The response comes back from that model. The usage record in your dashboard shows the model name, the tokens the provider reported, and the charge calculated at the rate card prices. You do not wonder whether a different tier answered. The curated line-up is the contract: the model you name is the model that answers.

BestRelay charges once per call after it ends, with no monthly fee and no minimum spend

You top up your prepaid balance through Stripe checkout. The amounts are fixed: $1.39, $6.94, $13.89, $27.78, $69.44, $138.89. There is no monthly fee, no subscription, and no minimum spend. Each call is charged once, after it ends, from the tokens the provider reports, at the rate card prices. Nothing is held in advance. Requests the provider rejects are recorded at zero. When the provider cannot be reached, nothing is charged. You can refund 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.

For example, you add $13.89 to a fresh account that already holds the $0.35 free credit. You run a day of work with GPT-6.1 Sol: 500,000 input tokens and 100,000 output tokens. At the BestRelay rate card, that day costs $0.22. Twenty-two such days cost $4.74. With 350,000 of the input read from cache, the day costs $0.16. The free credit alone covers about 1.6 million GPT-6.1 Sol input tokens. You watch the balance decrement call by call, and the usage log shows every model, every token count, every charge.

BestRelay connects Claude models only to Claude Code and Claude Desktop, and Gemini only through its native protocol

You configure your client with two settings: the address and a key created under API keys in the dashboard. BestRelay enforces client compatibility at the relay layer. 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. You do not need to guess which model works with which tool; the relay rejects mismatched pairs before they reach the provider.

You have an account limit of 120 requests a minute on one key, and 8 calls in flight on one account. An account can hold several keys, each can be voided at any time, and every call appears in the usage records with its model, tokens, and charge. 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. You see the privacy boundary in the records: your prompt text is retained briefly, the model output is not.

BestRelay referral commission moves into balance or withdraws to Alipay from $20.83

You share your referral link. When someone joins through it, you earn 18% of their top-ups. When they invite someone else, you earn 4% of that person's top-ups. Commission can be moved into the balance, or withdrawn to Alipay from $20.83. As an example, on an invited member's top-up of $13.89, the two commissions are $2.50 and $0.5556. You track these in the dashboard alongside your usage. The referral program does not change the rate card; it adds a separate credit stream that you control.

You compare the rate card to official list prices where they are published. For example, a day with Claude Sonnet 5.5 at 500,000 input and 100,000 output tokens costs $1.40 on BestRelay. Twenty-two 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. You see the numbers in your own usage log, not in a marketing table.

BestRelay keeps the line-up current by delisting retired model ids

You return to the dashboard next week. The list still shows twenty models, but a version number may have changed. When a provider retires a model id, BestRelay delists it. You do not send requests to deprecated endpoints. The curation is active, not static. You see DeepSeek V4 Pro at $1.53 input and $4.58 output with a cache read price of $0.0514. You see Kimi K3 at $1.39 input and $7.01 output with a cache read of $0.1389. You see GLM-5.3 at $0.625 input and $1.94 output with a cache read of $0.1389. Each entry is a living pointer to a provider model that exists today.

You choose a model, you send a request, you get an answer from that model. The charge matches the rate card. The record matches the provider's token count. The client rule matches the provider's requirement. BestRelay does not add features the sheet does not list, and it does not hide fees the sheet does not mention. Try BestRelay and see the curated line-up work at your desk.