GPT-6.1 Sol pricing via Vana API: $0.1389 per million input tokens

Who is on the listEN2026-11-05

§ One address, three protocols

Vana API is built around a simple arrangement: one address, three protocols, minimal setup. The address that answers API requests is https://api.vanaapi.com. The dashboard, where keys, balance and the usage log live, sits on the main site at https://vanaapi.com. You create a key, top up a prepaid balance, point your client at the address, and begin calling models.

Which protocol you use follows from the model and the client you already have:

GPT, Grok, DeepSeek, GLM, Kimi, Qwen, MiniMax and MiMo models use the OpenAI format at /v1/chat/completions, and work from Cursor, Cline, Roo Code, OpenCode, Cherry Studio, any OpenAI SDK or cURL.
GPT models also answer on /v1/responses, which is what Codex uses.
Claude models use the Anthropic Messages format at /v1/messages, and work from Claude Code or Claude Desktop, version 2.1.285 or later. They do not work from any other client, any SDK, cURL or a script.
Gemini models use the Gemini native protocol at /v1beta/models/..., and work from Gemini CLI. They are not OpenAI-compatible.

For the subject of this article, GPT-6.1 Sol, the route is short. It is a GPT model, so it speaks the OpenAI format, and it also answers on the responses path used by Codex. The rest of this walkthrough uses GPT-6.1 Sol as the worked example, with every sum labelled as an example and every rate taken from the published Vana API price list.

§ GPT-6.1 Sol pricing: a worked example

GPT-6.1 Sol, model id gpt-6.1-sol, is listed on Vana API at $0.1389 per million input tokens and $0.8333 per million output tokens. The official list price for the same model is $2 per million input tokens and $10 per million output tokens. The model id is what you place in your request, and the charge is applied per real token against your prepaid balance.

Here is the first worked example, and it is only an example. Say you send 1,000,000 input tokens and receive 1,000,000 output tokens in a single request:

Input: 1 million tokens at $0.1389 per million = $0.1389
Output: 1 million tokens at $0.8333 per million = $0.8333
Total for this example request: $0.1389 + $0.8333 = $0.9722

That is the arithmetic for this example. There is no monthly fee on the pay-as-you-go path, so the request above is charged at the published token rates, and every call is listed in the usage log in the dashboard, where you can review the token counts behind the figure.

Now a second example, again an example rather than a forecast of your own costs. Say you run this pattern every day: 10,000,000 input tokens and 2,000,000 output tokens across the day. Using the same published rates:

Input: 10 million tokens at $0.1389 per million = 10 × $0.1389 = $1.389
Output: 2 million tokens at $0.8333 per million = 2 × $0.8333 = $1.6666
Total for this example day: $1.389 + $1.6666 = $3.0556

Multiply that example day by 30 and you get 30 × $3.0556 = $91.668 for the month, all of it token-priced and all of it traceable in the usage log. The arithmetic is deliberately plain, because pay-as-you-go pricing should be something you can reproduce on a notepad before you spend anything.

§ Comparing with the list price

Run the same two examples against the official list price of $2 input and $10 output per million tokens and the gap becomes easy to read. For the first example request, the list price comes to 1 × $2 + 1 × $10 = $12. On Vana API the same request is $0.9722, a difference of $12 − $0.9722 = $11.0278.

For the second example day, the list price comes to 10 × $2 + 2 × $10 = $40, against $3.0556 on Vana API, a difference of $40 − $3.0556 = $36.9444. Across 30 identical example days, the list-price total would be 30 × $40 = $1,200, and the Vana API total would be 30 × $3.0556 = $91.668.

These are worked examples, and your own totals depend on the tokens you actually send and receive. The method, however, stays the same: the rate card is published per model and per direction, the usage log records the calls, and you can check the token counts behind any figure you see.

GPT-6.1 Sol also sits alongside the other GPT entries on the list, including GPT-5.6 Sol at $0.3472 input and $1.81 output per million tokens, and GPT-6 Astra at $0.6944 input and $3.54 output per million tokens. The practical question is not how one model compares with another, but what a given traffic pattern adds up to on the rate you will actually be billed. Vana API publishes the Vana API rate and the official list price for GPT-6.1 Sol side by side, so the comparison can be made with the same two numbers.

§ Setting up GPT-6.1 Sol with minimal setup

The setup path for GPT-6.1 Sol is short. Open https://vanaapi.com, create a key in the dashboard, and top up your prepaid balance. Sign-up includes $0.10 of credit, which is enough to place a first small request and confirm the path works. The API itself is reached at https://api.vanaapi.com.

From there, pick the client that matches how you already work. Because GPT-6.1 Sol speaks the OpenAI format at /v1/chat/completions, you can use Cursor, Cline, Roo Code, OpenCode, Cherry Studio, any OpenAI SDK or cURL. If you prefer the responses path, GPT models also answer on /v1/responses, which is what Codex uses. In both cases the base address is the same one, and the model id is gpt-6.1-sol.

Two details are worth keeping straight when one address serves several model families. Claude models require Claude Code or Claude Desktop, version 2.1.285 or later, because they speak the Anthropic Messages format at /v1/messages themselves. Gemini models require Gemini CLI and the Gemini native protocol at /v1beta/models/..., and that protocol is not OpenAI-compatible. If you later add one of those families, the client and the protocol change, while the address and the key stay as they are.

Because billing is prepaid and charged per real token, there is no plan to size before you begin. You can start with a small balance, watch the usage log, and adjust from there. The monthly plan for GPT is a separate option: up to $2,750 / 30 days for $103.61 for 30 days, at most $820 a week, GPT models only. If your GPT usage is steady and predictable, that plan places a ceiling on spend; if it varies, pay-as-you-go keeps the cost tied to what you use.

§ Balance, plans, referrals and refunds

Vana API bills against a prepaid balance with no monthly fee on the standard path. Every call appears in the usage log in the dashboard, and keys are created in the same place, so one page covers both activity and access. The dashboard is at https://vanaapi.com, while the address that answers API requests remains https://api.vanaapi.com.

Three monthly plans are available, each scoped to one model family. The Claude plans are Claude up to $450 / 30 days for $64.72 for 30 days, at most $140 a week, Claude models only, and Claude up to $1,050 / 30 days for $153.61 for 30 days, at most $320 a week, Claude models only. The GPT plan is GPT up to $2,750 / 30 days for $103.61 for 30 days, at most $820 a week, GPT models only. Each plan states its own limit and its own weekly cap, so a plan can be matched to a budget.

Referrals are handled on the same prepaid basis. You earn 14% of what the people you invite pay, and 5% one level further. Earnings can be withdrawn from $13.89, which marks the starting point for a first withdrawal.

Refunds are defined in a narrow, specific way: unused balance can be refunded within 3 days of the top-up, minus what was used and the payment processor's fee. Keeping the top-up and the usage log visible makes that calculation straightforward, since you can see the amount added and the amount consumed.

Put the pieces together and the case for Vana API is mostly arithmetic. One address at https://api.vanaapi.com, three protocols behind it, and a key created in the dashboard at https://vanaapi.com. GPT-6.1 Sol at $0.1389 input and $0.8333 output per million tokens, with the official list price shown alongside it. In the worked example above, a month of 10 million input tokens and 2 million output tokens per day came to $91.668 on Vana API and $1,200 at list price. Those figures describe an example pattern, and your own totals follow your own traffic, but the method is the same: price the tokens, check the log, and pay for what you use.