Riven Pricing Brain

The single source of truth for what things cost: model rates, tier multipliers, plan allowances, and the routing economics behind them.

Prices appear in a lot of places — the pricing page, Console, quota responses, the model catalog, routing decisions. Pricing Brain exists so all of them read from one definition instead of drifting apart.

What it holds#

• Upstream rates — per-model input and output cost for every cloud provider, and modelled cost per token for on-prem pools.
• Tier multipliers — the factor converting tokens used into allowance spent for each Riven tier.
• Plan definitions — allowance, included models, burst limits, and overage policy per plan.
• Routing economics — the cost signal Ultra Router and Spot Broker weigh against latency and quality.

How metering works#

Each request consumes *tokens used × the tier's multiplier* from your monthly allowance. A 1,000-token answer on a 1× tier costs 1,000 allowance tokens; the same answer on a 13.3× tier costs about 13,300. Routed aliases are billed at whichever lane actually served the message, so easy messages stay cheap without configuration.

Pay-as-you-go keys are metered the same way against a base rate of $3.00 per million tokens: effective price = base rate × tier multiplier. Requests pinned to a specific frontier model are billed at that vendor's list rate instead of a multiplier. The full multiplier table is in Models & usage.

Reading prices programmatically#

# Tier names, ordering, and multipliers — no auth required
curl https://api.rivenai.io/v1/catalog?view=tiers