// access tiers
Hold the token. Use the network.
There is no subscription and no card on file. Your live $MODEL balance — held in your own wallet, never deposited with us — sets your monthly usage allowance and rate limit.
// entry
Explorer
100K
$MODEL held to qualify
- $2 usage allowance / month
- 10 requests / minute
- Full model catalogue + agents
- Unlimited API keys, per-request activity
// build
Builder
1M
$MODEL held to qualify
- $15 usage allowance / month
- 30 requests / minute
- Full model catalogue + agents
- Unlimited API keys, per-request activity
// scale
Operator
5M
$MODEL held to qualify
- $100 usage allowance / month
- 90 requests / minute
- Full model catalogue + agents
- Unlimited API keys, per-request activity
// enterprise
Institutional
20M
$MODEL held to qualify
- $600 usage allowance / month
- 300 requests / minute
- Full model catalogue + agents
- Unlimited API keys, per-request activity
// how metering works
Every rule, in plain text.
QualifyingMinimum 100K $MODEL held in a wallet you control
Recognised capBalances above 30M $MODEL are clamped for tier purposes
Metering(input tokens ÷ 1M × input rate) + (output tokens ÷ 1M × output rate)
ResetAllowance resets at the start of each month, 00:00 UTC. No rollover.
Rate limitsEnforced per key, per minute. Over-limit returns 429 with Retry-After.
CustodyNothing is deposited, escrowed or transferred. Balances are read on-chain.
DowngradeBalance below the tier threshold steps access down at the next balance read.
Per-model token rates are published live on the models page. Call details, streaming and error handling are documented in the API docs.