Cursor on Vana API: pay-as-you-go with $0.10 credit to start

Money mattersEN2026-11-03

§ One address, three protocols, minimal setup

Vana API gives you a single address for model traffic: https://api.vanaapi.com. That address answers API requests, and behind it sit three protocols, so the same endpoint can serve different clients and different model families without a separate account for each. The Anthropic Messages protocol and the Gemini native protocol are served from the same place as the OpenAI format, which means your base URL stays constant while your tooling changes.

The dashboard lives somewhere else, on the main site at https://vanaapi.com. That is where you create keys, check your balance and read your usage log. Keeping the request address and the account area separate makes the setup small: one base URL in your editor, one key generated in the dashboard, and a prepaid balance ready when you are.

Because all three protocols share one address, moving a workflow does not mean rebuilding your infrastructure. You can point one tool at the OpenAI format, another at the Anthropic Messages protocol and a third at the Gemini native protocol, and every request still travels to the same Vana API endpoint. There is no monthly fee attached to that arrangement. You top up a balance, Vana API charges per real token, and you can see each call in the usage log afterwards.

For teams that want a predictable ceiling rather than a subscription, Vana API also offers a monthly plan. It is described later in this article, alongside the pay-as-you-go option, so you can compare a fixed 30-day price with per-token spending on your own traffic.

§ Getting Cursor onto the OpenAI format

Cursor speaks the OpenAI format, so it connects to Vana API through /v1/chat/completions on the same single address. Set the base URL to https://api.vanaapi.com, paste in a key you created in the dashboard, and choose a model from the list below. Cursor is not limited to one family: it can also reach Grok, DeepSeek, GLM, Kimi, Qwen, MiniMax and MiMo models, because they all answer in the OpenAI format.

GPT models answer on a second route as well, /v1/responses, which is the endpoint Codex uses. The same models therefore work from Cursor, Cline, Roo Code, OpenCode, Cherry Studio, any OpenAI SDK or cURL, depending on which client you prefer. If you script your requests directly, cURL and the OpenAI SDK take the same base URL and the same key, so there is nothing extra to register.

Here is a short checklist for a Cursor setup:

Create a key in the dashboard at https://vanaapi.com.
Add https://api.vanaapi.com as the base URL in Cursor.
Insert the key and pick a model id such as gpt-5.6-sol, deepseek-v4-pro or qwen3.8-max.
Send a test request, then confirm it in the usage log.

That is the whole configuration path. There is no account pool to arrange and no per-client signup. Vana API keeps the protocol details in one place, and Cursor keeps the interface you already know. The practical result is that switching a Cursor session to a cheaper model is a model id change rather than an integration project.

§ What the models cost per million tokens

Vana API prices every model per million tokens, split into input and output. The numbers below are exact, taken from the Vana API price list, and they are what your prepaid balance is charged against.

GPT-5.6 Sol (model id gpt-5.6-sol): Vana API $0.3472 / $1.81; official list price $4 / $20.
GPT-6 Astra (model id gpt-6-astra): Vana API $0.6944 / $3.54; official list price $10 / $50.
GPT-6.1 Sol (model id gpt-6.1-sol): Vana API $0.1389 / $0.8333; official list price $2 / $10.
Grok 4.7 (model id grok-4.7): Vana API $0.4861 / $1.53.
DeepSeek V4 Flash (model id deepseek-v4-flash): Vana API $0.2778 / $1.04.
DeepSeek V4 Pro (model id deepseek-v4-pro): Vana API $1.11 / $3.33.
GLM-5.3 (model id glm-5.3): Vana API $0.5556 / $1.67.
Kimi K3 (model id kimi-k3): Vana API $1.04 / $5.21.
Qwen3.7 Plus (model id qwen3.7-plus): Vana API $0.1278 / $0.4167.
Qwen3.8 Max (model id qwen3.8-max): Vana API $0.6944 / $2.08.
MiniMax M3 (model id MiniMax-M3): Vana API $0.1389 / $0.5556.
MiMo V2.6 Pro (model id MiMo-V2.6-Pro): Vana API $0.2083 / $0.3472.

The spread is wide enough that a single default model is rarely the right answer for every task. Qwen3.7 Plus and MiniMax M3 sit at the lower end of this list, GPT-5.6 Sol sits in the middle, and DeepSeek V4 Pro and Kimi K3 sit higher. Since Cursor can call all of them through the same OpenAI format, you can match the model to the job without changing your endpoint or your key.

For an example, say you run a Cursor session that sends 1 million input tokens and 1 million output tokens through GPT-5.6 Sol. Using the Vana API figures of $0.3472 and $1.81, that example costs $0.3472 + $1.81 = $2.1572. The same example on GPT-6.1 Sol, at $0.1389 and $0.8333, costs $0.1389 + $0.8333 = $0.9722. The example is simple arithmetic on the published rates, and it shows why per-token pricing rewards a deliberate model choice.

§ Pay-as-you-go billing, keys and the usage log

Vana API uses a prepaid balance. You add funds, and every call draws down that balance according to the real token counts of the request. There is no monthly fee on this route, so an occasional heavy day and a quiet month both cost only what you actually used. For example, if you only make a handful of Cursor requests in a given week, you are not paying for a plan you barely touch.

Keys are created in the dashboard, and every call is listed in the usage log there. That log is the record of what was spent and when, which makes it straightforward to check a bill against your own activity rather than against a rounded subscription figure. The same dashboard shows your remaining balance, so you can see how much runway is left before topping up again.

New accounts receive $0.10 of credit on sign-up. It is a small amount, but it is enough to place a real request through Vana API and confirm that your key, base URL and model id are all working before you commit any larger spend.

Refunds are handled on unused balance. If you top up and change your mind, the unused balance can be refunded within 3 days of the top-up, minus what was used and the payment processor's fee. The window is short, so it pays to check your balance and your planned traffic soon after adding funds.

§ The monthly plan and referral earnings

If a fixed ceiling suits your usage better than per-token charging, Vana API lists a monthly plan for GPT models only: GPT up to $4,000 / 30 days: $103.61 for 30 days, at most $1,200 a week. The plan is priced for 30 days, capped at $1,200 a week, and limited to GPT models. It is the option to consider when you already know your GPT volume is high and steady enough to justify a set figure.

For everyone else, pay-as-you-go through Vana API remains the flexible path. You keep the full model list available to Cursor, you pay per real token, and you avoid a monthly commitment you may not need. The two options can be compared directly: the plan gives you a known 30-day price against a $4,000 cap, while prepaid balance tracks whatever you genuinely send.

Vana API also pays for referrals. You receive 14% of what people you invite pay, and 5% one level further. Earnings can be withdrawn from $13.89. The percentages are calculated on what your invited users actually pay, so the scheme follows real activity rather than sign-ups alone.

Put together, the offer is straightforward: one address at https://api.vanaapi.com, three protocols behind it, a dashboard for keys, balance and usage, and pricing that lets you start small and scale with your own traffic. Cursor connects through the OpenAI format with minimal setup, GPT and the wider model list are available on the same endpoint, and Vana API charges per real token with no monthly fee on the pay-as-you-go route.