What each plan costs, what an API call is billed as, and where the platform fee applies.
Two ways to pay. A plan is a monthly subscription for Chat, Panel and Studio. The API is billed on what each call actually used, against an organization balance. Pricing is where you subscribe.
Annual billing shows as a monthly figure: Plus $16.67, Pro $37.50, Teams $25 per seat; billed annually at $200, $450 and $300 per seat. Annual is billed once, up front, at ten times the monthly price, which is two months free: Plus saves $40, Pro saves $90 and Teams saves $60 per seat a year.
Teams starts at 3 seats. The price per seat is the same at any size. Seats can be added or removed at any time, and changes are prorated for the rest of the period.
Subscriptions renew until cancelled. Cancelling takes effect at the end of the period you paid for. Terms of Service sections 04 and 06 have the details.
A plan is sized in deliberation time. A deliberation runs several models, and its cost depends on how hard the question is and how much the panel disagrees. Plans include about 20 hours of deliberation time a month on Plus, about 50 on Pro. Every answer carries a receipt showing what it used.
Each plan reaches engines up to a ceiling, and the ceilings stack: Plus reaches everything Free does, Pro everything Plus does. Signed out reaches Express. Signed in on Free reaches Foundation. Plus and Pro reach Frontier. You pick Express, Foundation or Frontier per question, up to that ceiling. Einstein Mode ยท QRUM:EINS is a Mode on Pro.
Turbo credits top up your account without changing your plan: $10, $25 or $100, one-time; credits do not expire. They are used as you go and have no cash value.
Every Mode bills three parts: the engines at the price their provider charges, the platform fee on that, and the Mode's fee.
The published per-tier rate is the Mode's fee only, in dollars per 1M tokens processed in the call (all seats and the deliberation), input and output. Output is priced at four times input. GET /v1/models shows each rate per tier. The rate is for planning. The bill is actual usage, so the engines and the deliberation of your call decide it.
POST /v1/estimate quotes the whole call for your prompt: engines, platform fee and the Mode's fee. It is always free. GET /v1/models returns the exact figures your key is entitled to, including any negotiated rate, and takes precedence over the published figures for your organization.
The fee is a percentage of what each seat's source billed for that seat.
| Seat ran on | Fee rate |
|---|---|
| Quorum's keys | 5% |
| Your own key (bring your own keys) | 0.5% |
The rate is set per source, and the two rates above are the defaults. The fee is on API calls only, on the source's real billed cost; never on plan usage. Each seat's source, real cost, fee rate and fee appear on the receipt. The fee is never charged on the Mode's fee.
Worked example. A seat whose source billed $0.04 pays a fee of $0.002 on Quorum's keys (5% of $0.04) and $0.0002 on your own key (0.5% of $0.04). The fee is added to the bill beside the engine cost and the Mode's fee.
An API-key call cannot bill more than its ceiling, which is the higher of $2.00 and the price of the Mode being called, unless the API key carries its own ceiling. Only Quorum sets a key's own ceiling. A request can set a lower ceiling for one call with quorum.max_cost_usd over REST, or the max_cost_usd argument of deliberate over MCP. Calls on your plan, including connected apps, pay real cost and have no per-call ceiling. When a call goes over the ceiling you still get the whole answer. A call that goes over its ceiling is charged in full: the engines, the platform fee and the Mode's fee. The amount over the ceiling is then credited back as a cap credit, so you pay the ceiling. The response carries quorum.capped: true and the credit in quorum.cap_credit_usd, and the receipt shows it as a cap credit line.