GPT-6.1 Sol on BestRelay: API relay pricing, read line by line

BestRelay forwards each request under the model name you wrote

BestRelay is an independent API relay, and the promise it makes is plain: each request is forwarded to the model provider under the model name the caller wrote, so the model named is the model that answers, and nothing is quietly swapped for a cheaper tier. That single rule is the reason the rest of the rate card can be read without suspicion. BestRelay lists a curated line-up of 20 public models, each under its own name, and each with four prices: input, output, cache write and cache read. Take GPT-6.1 Sol. Its line reads $0.2083 for input per 1,000,000 tokens, $1.11 for output, and $0.0431 for cache read. The official list price for the same three entries is $2 input, $10 output and $0.1 cache read. No cache write price is listed for GPT-6.1 Sol on the card, and BestRelay leaves the gap as it is rather than filling it with a guess. The card is the card: every figure can be read before a single token is sent, and the model you chose is the model that works on your behalf.

Reading the same four columns across the curated line-up

The discipline holds for every entry, which makes comparison a matter of reading rather than inference. A few lines from the BestRelay rate card, each per 1,000,000 tokens:

• Claude Fable 5: input $4.17, output $20.97, cache write $5.14, cache read $0.3472.
• Claude Opus 5.5: input $1.74, output $8.13, cache write $2.01, cache read $0.0778.
• Claude Sonnet 5.5: input $1.39, output $7.01, cache write $1.74, cache read $0.0944.
• Gemini 3.1 Pro: input $1.74, output $9.93, cache write $1.74, cache read $0.1389.
• DeepSeek V4 Pro: input $1.53, output $4.58, cache read $0.0514.
• Kimi K3: input $1.39, output $7.01, cache read $0.1389.
• Qwen3.8 Max: input $0.8333, output $2.57, cache read $0.1014.
• Grok 4.7: input $0.625, output $1.94, cache read $0.1389.
• GLM-5.3: input $0.625, output $1.94, cache read $0.1389.
• GPT-5.6 Sol: input $0.4167, output $2.29, cache read $0.0333.
• GPT-6 Astra: input $0.7639, output $3.96, cache read $0.0792.

At the other end of the sheet, Gemini 2.5 Flash is listed at $0.2778 input, $2.08 output, $0.2778 cache write and $0.0264 cache read, while Claude Haiku 4.5 sits at $0.4167 input, $2.01 output, $0.4861 cache write and $0.0333 cache read. Some lines carry four figures and some carry three, and BestRelay prints exactly what the provider reports for each. The useful habit is to read across the row you intend to use: cheap input does not imply cheap output, and a low cache read rate rewards a workflow that reuses context rather than resending it. Because BestRelay forwards each request under the name you wrote, the row you read is the row you pay for.

How far a million tokens goes in an ordinary working day

The card becomes easier to judge once it is applied to a familiar shape of work. For example, a day of 500,000 input tokens and 100,000 output tokens on GPT-6.1 Sol 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; the difference on the day is $1.78 and on the 22 days $39.26. The free credit alone covers about 1.6 million GPT-6.1 Sol input tokens, which is enough to see the pattern with your own material rather than a sample. Another line reads similarly. Claude Sonnet 5.5, as an example: a day of 500,000 input and 100,000 output tokens costs $1.40 on BestRelay, 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. Caching is the lever both lines share, and BestRelay prices it openly in every row that carries a cache read figure.

How the balance moves, and what happens when a call fails

On BestRelay, billing follows the call rather than the calendar. The account holds a prepaid pay-as-you-go balance: no monthly fee, no subscription, 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, so a quiet week costs what it used and no more. Failed calls are recorded at zero, and when the provider cannot be reached, nothing is charged at all. Each new account opens with $0.35 of credit. Top-ups run through Stripe checkout in these amounts: $1.39, $6.94, $13.89, $27.78, $69.44 and $138.89, so the balance can be topped in small steps or in larger ones depending on how much you expect to run. 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. Referrals pay 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. As an example, on an invited member's top-up of $13.89, the two commissions are $2.50 and $0.5556.

Two settings, clear client rules, and records you can read later

Connecting to BestRelay takes two settings: the address, and a key created under API keys in the dashboard. The docs carry the details, and neither is printed here. The client rules are equally specific. 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. Tools such as Codex pair 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. On one key, the limit is 120 requests a minute, and one account can have 8 calls in flight. 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, so a rate card can be checked against real usage rather than trusted on faith. On privacy, 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. When a provider retires a model id, BestRelay delists it, keeping the line-up curated. For example, if you start with GPT-6.1 Sol at $0.2083 per 1,000,000 input tokens, you can watch each charge appear in the records and try BestRelay yourself.