GPT-5.6 Sol pay-as-you-go on Vana API from $0.3472
§ One address, three protocols, minimal setup
Vana API runs on a single address, https://api.vanaapi.com, and that is the only endpoint you need for API requests. The dashboard sits separately at https://vanaapi.com, where you create keys, monitor your balance, and read the usage log. This split keeps the request path simple while giving you a dedicated place to manage everything administrative.
Three protocols share that one address. Claude models use the Anthropic Messages protocol at /v1/messages, Gemini models use the Gemini native protocol at /v1beta/models/..., and GPT, Grok, DeepSeek, GLM, Kimi, Qwen, MiniMax, and MiMo models use the OpenAI format at /v1/chat/completions. GPT models also respond on /v1/responses. Because all of these run from the same URL, you do not need to juggle multiple endpoints or remember which service lives where.
The setup is deliberately minimal. You point your client at https://api.vanaapi.com, add the key from the dashboard, and start sending requests. There is no separate onboarding flow for each model family, no extra configuration files, and no need to route traffic through different hosts. Vana API keeps the plumbing quiet so you can focus on the work.
The address https://api.vanaapi.com answers API requests only. It does not host a web interface, which keeps the request path lean. When you need to manage keys or read the usage log, you go to https://vanaapi.com. This division means the API endpoint stays focused on serving model responses while the dashboard handles account management.
§ Choosing the right client for each model family
The protocol determines which client you use, and Vana API documents this mapping clearly. For Claude models, you work from the official Claude Code or Claude Desktop, version 2.1.285 or later, which speak Anthropic Messages themselves. For Gemini models, you use Gemini CLI over the Gemini native protocol. For GPT-5.6 Sol and the other OpenAI-format models, you can use Cursor, Cline, Roo Code, OpenCode, Cherry Studio, any OpenAI SDK, or cURL.
This structure means Vana API fits into the tools you already use. If you already run Cursor for OpenAI-format models, the same setup works for GPT-5.6 Sol on Vana API. If you use Claude Desktop for Claude models, that client handles the protocol details. You do not need to install new software or learn a new interface for each model.
For example, when you run GPT-5.6 Sol through Cursor, the client sends OpenAI-format requests to /v1/chat/completions at the single address. When you use Claude Desktop with a Claude model, it sends Anthropic Messages to /v1/messages. When you use Gemini CLI, it speaks the Gemini native protocol. Vana API routes each request to the right model based on the path and the key.
§ Reading the usage log after a GPT-5.6 Sol session
After a session with GPT-5.6 Sol, open the dashboard at https://vanaapi.com and read the usage log. Every call is listed there, showing the model, the tokens consumed, and the charge. Vana API bills from a prepaid balance, charged per real token, with no monthly fee. GPT-5.6 Sol costs $0.3472 per million input tokens and $1.81 per million output tokens on Vana API, against an official list price of $4 / $20.
Say you run this every day. Each entry in the log adds up precisely, so you can see exactly what a session cost without waiting for an invoice. The log is the single source of truth for spending on Vana API, and it updates as calls complete. This transparency makes it easy to match usage to projects or clients.
The usage log shows each call with its timestamp, model ID, and token counts. For GPT-5.6 Sol, you will see the model id gpt-5.6-sol in the entries. Because billing is per real token, the log reflects actual consumption rather than estimates. You can export or review these entries to understand spending patterns over time.
Reading the log after a session gives you a clear picture of cost. For GPT-5.6 Sol, the input price of $0.3472 per million tokens and output price of $1.81 per million tokens apply to every entry. The official list price for the same model is $4 / $20, so the log entries reflect the actual Vana API rates rather than the list rates.
§ Billing, keys, and small details that matter
Keys are created in the dashboard, and each key can be tracked in the usage log. New accounts receive $0.10 of credit on sign-up, which is enough to test the connection and read the first log entries. Unused balance can be refunded within 3 days of the top-up, minus what was used and the payment processor's fee.
If you invite others to Vana API, the referral program pays 14% of what they spend, and 5% one level further. Earnings can be withdrawn from $13.89. These details live in the dashboard alongside the balance and usage log, so everything about your account is in one place.
The refund window of 3 days applies to the unused portion of a top-up. If you add funds and decide not to use them, you can request a refund of the remainder after deducting used amounts and the processor fee. The referral earnings accumulate in the dashboard and become withdrawable once they reach $13.89.
The dashboard at https://vanaapi.com is where all account activity lives. You create API keys there, set up your prepaid balance, and open the usage log to review every call. Because Vana API charges per real token with no monthly fee, the balance decreases only when requests are processed. This pay-as-you-go approach keeps costs aligned with actual use.
§ Monthly plans for predictable workloads
For steady usage, Vana API offers monthly plans alongside the pay-as-you-go balance. The Claude up to $1,000 / 30 days plan costs $64.72 for 30 days with at most $320 a week, and the Claude up to $2,400 / 30 days plan costs $153.61 for 30 days with at most $750 a week. The GPT up to $4,000 / 30 days plan costs $103.61 for 30 days with at most $1,200 a week. These plans are limited to their respective model families as listed.
If your usage varies, the prepaid balance remains the flexible choice. You top up when needed, watch the usage log, and only pay for the tokens you actually consume. Vana API keeps both paths available, so you can switch between pay-as-you-go and a monthly plan as your needs change.
Choosing between pay-as-you-go and a monthly plan depends on your volume. The pay-as-you-go model on Vana API lets you start with the $0.10 sign-up credit and scale up only when you need to. Monthly plans suit predictable workloads where a fixed spend cap is helpful. Both options use the same address and the same usage log.
Monthly plans on Vana API provide a spend cap for users who prefer predictable billing. The limits are explicit: for example, the GPT plan allows up to $4,000 over 30 days, with at most $1,200 in any single week. If you exceed the plan scope, the usage log still records every token, and you can adjust your plan or return to pay-as-you-go.