Pricing & billing
boldrouter is prepaid and metered per token. You buy credits, and each request decrements your balance at the model's price. No seats, no minimums.
Prepaid credits
Every account holds one credit balance shared across all providers. Top up to start making requests; when the balance hits zero the gateway returns 402 insufficient_quota until you top up again - so you can never be billed beyond what you've prepaid.
Per-token pricing
Each request is metered on its token usage (taken from the provider's own counts when reported, estimated otherwise) and priced per the route that served it. Credits are 1:1 with provider spend - there's no per-token markup; the platform's fee is taken once when you buy credits (below).
| Field | Type | Description |
|---|---|---|
| Platform key | default | The provider's list price, exactly as shown in the model catalog - your credits buy provider tokens at cost. |
| Your own key (BYOK) | reduced | Just a 5.5% platform fee on the provider's list price - you pay the provider directly for the tokens. See Bring your own key. |
Cached prompt tokens are billed at a steep discount to the input rate - see Prompt caching.
Long-context (tiered) pricing
A few long-context models charge a higher rate once a single request's prompt crosses a token threshold - and the provider applies that rate to the whole request, not just the tokens past the threshold. boldrouter mirrors this exactly: it compares the request's prompt-token count to the model's threshold and bills the entire request at the matching tier.
| Field | Type | Description |
|---|---|---|
| google/gemini-2.5-pro | > 200K tokens | $1.25 / $10 per 1M at or below 200K prompt tokens; $2.50 / $15 above. |
| gemini-3.1-pro-preview | > 200K tokens | $2 / $12 per 1M at or below 200K prompt tokens; $4 / $18 above. Served as a Bold Frontier route. |
| minimax/minimax-m3 | > 512K tokens | $0.30 / $1.20 per 1M at or below 512K prompt tokens; $0.60 / $2.40 above. |
| openai/gpt-5.6-sol | > 272K tokens | $4 / $20 per 1M at or below 272K prompt tokens; $8 / $30 above. |
| openai/gpt-5.6-terra | > 272K tokens | $2 / $12 per 1M at or below 272K prompt tokens; $4 / $18 above. |
| openai/gpt-5.6-luna | > 272K tokens | $0.20 / $1.20 per 1M at or below 272K prompt tokens; $0.40 / $1.80 above. |
| openai/gpt-5.5 | > 272K tokens | $5 / $30 per 1M at or below 272K prompt tokens; $10 / $45 above. |
| openai/gpt-5.4 | > 272K tokens | $2.50 / $15 per 1M at or below 272K prompt tokens; $5 / $22.50 above. |
Off-peak pricing
Some providers charge less outside their busiest hours. boldrouter bills each request at the rate in force when it's made, so you get the off-peak discount automatically. Prices in the model catalog are the peak rates.
| Field | Type | Description |
|---|---|---|
| DeepSeek (all models) | 50% off | Peak hours are 01:00-04:00 and 06:00-10:00 UTC, Monday to Friday. Outside them, input, cached input and output tokens all cost 50% of the peak rate. |
Platform fee (on top-ups)
boldrouter's entire fee is a 5.5% platform fee applied when you add credits - not per token. The amount you choose lands in your balance as real provider spend; the fee is added to what you pay. For example, topping up $100.00 in credits:
| Field | Type | Description |
|---|---|---|
| Credit added | balance | $100.00 |
| Platform fee (5.5%) | fee | $5.50 |
| You pay | total | $105.50 |
The fee appears as its own line item on your Stripe receipt. Manage top-ups on the Billing page, where you can pick a preset or enter any amount from $5 to $10,000.
Auto top-up
To keep requests from stopping when the balance runs out, turn on auto top-up on the Billing page: save a card, choose a balance threshold and an amount, and boldrouter adds that amount of credit whenever the balance drops under the threshold.
- Each automatic top-up is charged like a manual one: the chosen credit plus the 5.5% fee, with a Stripe invoice and a receipt by email.
- At most three automatic top-ups run in any 24 hours, so a runaway integration can't keep charging the card.
- If a charge fails, the account owner gets an email and the next attempt waits two hours. After three failed charges in a row, auto top-up turns itself off until a working card is added and it's turned back on.
- Card details stay with Stripe; boldrouter stores only the card type and its last four digits. Removing the card turns auto top-up off.
At a glance
- Credit top-ups: +5.5% platform fee - this is boldrouter's only charge.
- Platform-key tokens: provider list price, no markup (credits are 1:1 with provider spend), including off-peak discounts where the provider offers them.
- BYOK tokens: 5.5% of provider list price (you pay the provider directly for the tokens).
- Prepaid & capped: spend can't exceed your balance; per-key monthly caps available.