One Agentic

Pricing Variables v2

Credit-model config — drives Pricing Assumptions v2
🔒  Internal Use Only  ·  Link the config file once — then “Save & apply” writes directly to pricing_config.json
given (external fact) assumption (internal modelling decision) observed (production data) Grey = derived / calculated
→ View Pricing Assumptions v2

LLM API Pricing

Raw model cost per 1M tokens — everything else derives from these.

ParameterTypeValueUnitNotes
Model name given Displayed in pricing table notes
Input token price given $ / 1M tokens — standard prompt & context
Cached input token price given $ / 1M tokens Prompt caching — reused system prompts & templates
Output token price given $ / 1M tokens — generated response
Revenue markup multiplier assumption × Applied to raw API cost to set customer price
Implied gross margin % Formula: 1 − (1 ÷ markup)

Credit Model & Neo Interactions Session

1 credit = creditUnit tokens (reference unit for billing). creditRaw is derived automatically — it adjusts whenever token prices change so the avg Neo session always costs the target number of credits.

Credit Unit

ParameterTypeValueNotes
Credit unit size assumption tokens Reference token count per credit
Target credit multiplier assumption Stable ratio — creditRaw derives from this
Credit raw cost derived = avgSessionCost ÷ neoTargetCredits — auto-updates with prices
Credit at markup creditRaw × markup

Standard Session — Neo Interactions (avg observed)

Token TypeTypeTokensCost (raw)
Fresh input observed
Cached input observed
Output observed
Avg session cost (raw)
Session price at market

Input tokens account for ~86% of session cost. Cached tokens reduce cost significantly — variance in fresh input is the primary cost driver.

Session Distribution — Planning Scenarios (100 runs)

Token typeTypeMinStd dev (σ)Max
Input tokens observed
Cached input observed
Output tokens observed

Min / Max bound the Minimum and Maximum planning scenarios. Std dev (σ) drives the Conservative (+1σ) scenario for all three token types.

Tier Design

Price = credits × creditMrk × (1 − discount). Set credits and discount — price derives automatically and tracks token price changes.

TierTypeCredits / moDiscount %Price / moMargin @ 100%
Micro assumption
Seed assumption
Mid assumption
Established assumption

Overage rate = creditMrk (/cr) — always above any tier’s implicit credit price, preserving upgrade incentive.

TAM Inputs

All ARR estimates use tier prices as the revenue unit — they update automatically when token prices or discounts change.

Early-Stage VC — Primary Segment

ParameterTypeValueNotes
Base fund count assumption funds Mid-scenario addressable early-stage VC funds
Optimistic fund count assumption funds High-scenario fund count

Late-Stage VC Segment

ParameterTypeValueNotes
Base fund count assumption funds Global late-stage / multi-stage VC funds — base scenario
Optimistic fund count assumption funds Broader global count — optimistic scenario
Base ARR (Established tier) lBase × Established price × 12
Optimistic ARR (Established tier) lOptimistic × Established price × 12

Late-stage funds mapped to Established tier — larger teams, higher deal volume.

Angel Investor Segment

ParameterTypeValueNotes
Conservative angel count assumption US early-adopter active angels
Base angel count assumption ACA active angel estimate (US)
Optimistic angel count assumption Broader active angel universe (US)

Angels are assumed to land on the Micro tier — solo investors with lower session volumes.

Excluded Segments — Ballpark Only

SegmentTypeCountBase ARR (Mid)Opt. ARR (Estab.)
Multi-stage VC assumption
PE / growth assumption
Family offices assumption
Corporate VC assumption
Investment banking assumption
Corporate development assumption
Total excluded ARR

Base uses Mid tier as ARPU. Optimistic uses Established tier. Not included in the primary VC TAM above.