Billing
Freemium access
Some models are free to use while your workspace keeps a minimum wallet balance. Here is exactly how the freemium tier works, the balance you need to keep, and what happens if you dip below it.
How freemium works
A model tagged Freemium is free to call — your wallet is not charged — while your workspace balance stays at or above the minimum (1,000 rupees by default). unoblox covers the provider’s cost for you while you are eligible. The minimum is only checked, never spent.
Eligibility is evaluated live on every request — there is no monthly allowance and no monthly cap. It depends only on your balance and your request rate right now. Freemium calls are subject to a fair-use limit of 10 requests per minute per workspace, at any balance (see Rate limits).
The minimum balance
The minimum is a balance you maintain, not a fee. Keep 1,000 rupees or more and freemium models are free — that 1,000 rupees stays in your wallet. If your balance falls below the minimum, freemium models simply bill at their normal published price, charged to your wallet as usual, until you top back up. (The default is 1,000 rupees; your account’s exact minimum shows on your billing page.)
| Wallet balance | Freemium call | You pay |
|---|---|---|
| 1,500 rupees | within 10/min | Free — balance stays at 1,500 rupees |
| 1,000 rupees (exactly the minimum) | within 10/min | Free |
| 800 rupees (below the minimum) | within 10/min | The model's published price |
| 1,500 rupees | 15 in one minute | First 10 free, the rest get HTTP 429 |
Examples
Staying eligible. You keep 1,200 rupees in your wallet and call a Freemium model 500 times today, all within 10/min. You pay nothing — your balance is still 1,200 rupees tomorrow.
Dipping below. You then spend 300 rupees on a paid model, so your balance drops to 900 rupees. The next Freemium call is billed at that model’s published price (charged to your wallet) until you top back up to 1,000 rupees — nothing breaks, it just isn’t free while you are under the minimum.
Below the minimum. Every Freemium model has a published price, shown on its model page (for example Qwen3.8-27B). While your balance is at or above the minimum, you pay nothing. If it drops below, calls are billed at that published price until you top back up. The minimum does not change the rate limit: the freemium fair-use limit (10 requests per minute per workspace) applies at any balance. Only a key with paid access enabled is exempt from it.
Rate limits
The freemium fair-use limit is 10 requests per minute per workspace, whatever your wallet balance. Beyond it, requests return HTTP 429 (UB-GW-241) — wait a moment, pin a paid model, or enable paid access on the key (which bills at published prices with no freemium limit).
Free models also run 3 requests at a time per workspace. While that many free requests are still running, another returns HTTP 429 (UB-GW-313) with Retry-After: 2 — wait for one to finish, then retry. Nothing is charged for the refused request, and it does not count toward the per-minute limit. Paid models and keys with paid access have no such limit. The general per-key and per-workspace limiter is covered in Rate limits.
Good to know
- Promotional credit is ordinary wallet credit: it is billed against like any other balance, and it counts toward the minimum. It is not separate “free usage.”
- BYOK (bring your own provider key) is a separate route with a small platform fee; it does not use the freemium gate.
- A model’s freemium status and price can change — the model’s page always shows its current tag and price.
Get started — keep 1,000 rupees in your wallet and call any Freemium model free.