Pricing

UberLLM's pricing is deliberately simple and transparent.

No markup on inference

You pay the provider's token price for the endpoint that served your request — nothing added on top. When routing picks a cheaper provider (or a lower-priced seller offer), you pay that lower price. Every request is billed from the provider's reported token usage; when an upstream omits usage, UberLLM falls back to a tokenizer estimate (flagged in Dashboard → Activity).

You can see the exact per-endpoint prices for any model in uberllm.dev/models or GET https://api.uberllm.dev/api/v1/public/models.

Prepaid credits + 5% credit fee

Inference is prepaid. When you top up your balance, a 5% fee applies to the credit purchase — the fee is on the money-in, not on inference. Example:

(Credited = amount ÷ 1.05.) Minimum top-up is $5. Top-ups are handled via Stripe on Dashboard → Credits; your balance is drawn down per request at the provider's real price.

Spend controls

• Per-key spend limits — set a daily / monthly / lifetime cap when you create a key. Requests that would exceed it return 402 key_spend_cap_exceeded.
• Per-key rate limits — an optional requests-per-minute ceiling per key.
• Balance floor — when credits run out, requests return 402 insufficient_credits until you top up. Reservations for in-flight requests are released if the request aborts or fails.

See Errors for the exact 402 shapes.

BYOK (bring your own key)

If you route with your own provider key (BYOK), UberLLM does not touch your credit balance — cost_usd = 0 and no reserve/settle happens. The request is still logged so you keep unified usage analytics across curated, marketplace, and BYOK traffic. BYOK is ideal when you already have provider capacity or committed spend you want to use through one gateway. Configure BYOK under Dashboard → BYOK.

Sellers

If you sell capacity on UberLLM, your earnings accrue to a payable balance and are paid out separately — there is no platform markup taken from buyers on your offers. See Sellers.