A seat is $500 per month with unlimited runs, at an early-cohort rate locked for the life of the account. Producing one report costs us $1.61–$6.95 in API spend, measured on real runs — published below rather than hidden.
A seat is $500 per month for the early cohort, with that rate held for the life of the account. The email address you subscribe with becomes your API key: send a company name from it and the deliverables come back to it. There is no dashboard to learn.
Consultant seat
$500 / month
Early-cohort rate, locked for the life of the account.
We are onboarding the first cohort by hand. Join the list, or call 435-919-8036.
Measured on real runs, not estimated: 1.29M input and 1.17M output tokens per report across roughly 386-479 billable model calls. At current published API rates that is $1.61 per report on the cheapest capable chain and $6.95 on a stronger one.
| Model chain | $ / M in | $ / M out | Per report | Per 100 |
|---|---|---|---|---|
| MiniMax M3 | $0.27 | $1.08 | $1.61 | $161 |
| GLM 5.3 | $1.40 | $4.40 | $6.95 | $695 |
| Kimi K3 | $2.55 | $12.75 | $18.21 | $1,821 |
The obvious assumption is that simulating 40,200 customers means 40,200 language-model calls. It does not. Agent choices are computed by a multinomial-logit discrete-choice model running locally — the same class of model used in academic consumer-choice research — so the 2,046,200 decisions in the largest run cost effectively nothing.
The API bill is research, synthesis and page authoring: 386-479 calls, not two million. One role — analysis — accounts for 97% of output tokens, and output is priced 3–5× higher than input on every provider. Which model runs that one role decides essentially the entire bill.
Method: every outbound HTTP request and response was captured at the
transport layer and paired by sequence; only pairs answered
HTTP 200 were counted. Rate-limited retries were excluded
because they are an artefact of a free API tier and would not occur on a
paid plan. Token counts are derived from body size at 4 characters per
token and are accurate to roughly ±20%.
Early access
Unum11 is onboarding a small first cohort of consultants, brokers, and investors. Join the list and we'll reach out with your seat. Early users lock in early pricing — for good.