RelayVano API relay pricing for GPT-6 Astra refactoring work

Can GPT-6 Astra carry a long refactor across a large codebase?

Q: What are we actually offering here? A: A practical way to put GPT-6 Astra behind a single relay while you work through a refactor that spans many files. A refactor is rarely one prompt. You map a module, rewrite a service, rename an interface, update the callers, then run the tests again and fix what broke. Each of those steps can mean large prompt payloads, long replies and a lot of back-and-forth.

Q: Why GPT-6 Astra for that kind of work? A: It is a GPT model, and on RelayVano it is priced for sustained use rather than a seat-based arrangement. The fare is $0.7403 per million prompt tokens, $3.70 per million reply tokens and $0.0736 per million cache-hit tokens, in approximate US dollars. For comparison, the official list price is $10 / $50 / $1 cache hit. That gap matters when a single afternoon of refactoring sends millions of tokens through the same endpoint.

Q: Do I need a vendor subscription to keep using it? A: No. RelayVano sells a prepaid balance and charges per token. There is no monthly fee, and the balance never expires, so a heavy refactor week and a quiet week draw from the same account without a renewal decision in between.

Which clients can pair with GPT-6 Astra on RelayVano?

Q: Can I use the tools I already have? A: Yes, and the pairing rules are clear. GPT-6 Astra is a non-Claude, non-Gemini model, so it works with Codex, Cursor, Cherry Studio, Cline, OpenCode, OpenAI SDKs and cURL. If your refactor workflow lives in an OpenAI-compatible editor or script, you can point it at RelayVano and keep the model selection you already know.

Q: What about Claude and Gemini models? A: They follow their own paths, with no exceptions. Claude models work only in the official Claude Code and Claude Desktop clients, version 2.1.285 or newer. Gemini models use the native Gemini protocol only, in Gemini CLI, and they are not OpenAI-compatible. RelayVano shows you which client pairs with which model, so you do not have to guess whether a request will even leave the building.

Q: How many keys do I juggle? A: One. A single API key reaches every model through https://api.relayvano.com, including the OpenAI-compatible /v1 route, the Anthropic /v1/messages route and the native Gemini /v1beta route. For a long refactor, that means one credential to rotate, one balance to watch and one usage page to read.

What does a day of refactor traffic actually cost?

Q: Can you show a realistic shape? A: Here is an example, and it is only an example. Suppose you send 200 requests a day, each carrying 20,000 prompt tokens and producing 4,000 reply tokens. At GPT-6 Astra rates, the prompt side is 4,000,000 tokens, which is $0.7403 times 4, or about $2.96. The reply side is 800,000 tokens, which is $3.70 times 0.8, or about $2.96. Together, that is roughly $5.92 for the day, and roughly $177.60 across 30 such days.

Q: What changes that number? A: Cache hits. A refactor session often resends the same context, and cache-hit tokens on GPT-6 Astra cost $0.0736 per million, far below the prompt rate. Prompt tokens, reply tokens, cache stores and cache hits are all priced separately, so you can see which part of the traffic is carrying the bill.

Q: Is there a lighter option for the routine edits? A: If some steps are small, GPT-6.1 Sol on RelayVano is listed at $0.1958 prompt, $0.9792 reply and $0.0389 cache hit per million tokens. Many teams point the mechanical renames at that fare and reserve GPT-6 Astra for the structural rewrites that need more room to think.

What happens when a stream drops halfway through?

Q: I disconnect during a long refactor response. Is the work lost? A: No. If a caller disconnects mid-stream, the request is still finished and billed on the tokens really used. You are not paying for a response you never received in full, and you are not paying for tokens that were never generated either.

Q: Can I audit the charges? A: Every request shows up with its charge on the usage page. When a refactor month adds up, you can walk the bill request by request instead of comparing a total against a vague memory of how many prompts you sent.

Q: What if something looks wrong? A: Logged-in members write to us under Messages in the dashboard. The same place you top up is the place you ask a question, which keeps support inside the account you are already using.

How do top-ups, refunds and referrals work?

Q: How do I add credit? A: Top-ups go through Stripe checkout. Cards work everywhere, and WeChat Pay is available where your account supports it. Unspent top-ups can be refunded within 9 days, which makes a large prepaid balance easier to reason about when a refactor project scales back down.

Q: Is there anything waiting for me at signup? A: New accounts start with $0.40 of credit, enough to send a few requests and confirm your setup before you commit to a larger top-up.

Q: Can I share the service and earn something? A: People who sign up through your link generate a referral share of 16% of every top-up, and 4% one level further. Earnings can be withdrawn from $5.56 or moved into the balance, so referral income can also fund your own refactor traffic.

Why does RelayVano suit developers working worldwide?

Q: What makes this an international relay? A: RelayVano is an independent relay made for developers worldwide, and the whole site, dashboard and docs come in six languages: 中文, English, Español, हिन्दी, Português, Français. Prices show in yuan on the Chinese pages and in approximate dollars on the others, while the charge itself is made in yuan.

Q: What is the setup like? A: In an OpenAI-compatible client, set the base address to https://api.relayvano.com, add your RelayVano key, and choose GPT-6 Astra as the model. In Codex, Cursor, Cherry Studio, Cline, OpenCode, an OpenAI SDK or a cURL call, the same key and base address apply. Start with a small request, confirm the charge on the usage page, then let the refactor run.

Q: Why try it now? A: One key, six languages, per-token billing that never expires and a fare of $0.7403 prompt, $3.70 reply and $0.0736 cache hit for GPT-6 Astra. Sign up for RelayVano, add your first top-up, and give that long refactor a backend that charges you for what you actually use.