02
Network
- parable usage
- robinhood chain mainnet
- attested hourly
ROBINHOOD CHAINID 4663BLOCK56,089,844GAS0.369 GWEIBLOCK TIME100 MSUSDG SUPPLY648M USDGETH$2,477.67PARABLE REQUESTS3TOKENS2,280MODELS425ROBINHOOD CHAINID 4663BLOCK56,089,844GAS0.369 GWEIBLOCK TIME100 MSUSDG SUPPLY648M USDGETH$2,477.67PARABLE REQUESTS3TOKENS2,280MODELS425
since 9/6/2026, 1:32:22 PM live · refreshed 0s ago
Requests processed
3
every call, including failures
Tokens generated
1,950
2,280 processed in total
Models available
425
live catalog
Active API keys
0
keys with traffic
Latency p50
2,998 ms
p95 5,804 ms
Compute
14,762 PCU
estimated compute units
Top models · by requests
| Model | Requests | Tokens | p50 | Share |
|---|---|---|---|---|
| ~anthropic/claude-sonnet-latest | 1 | 36 | 2,184 ms | 33% |
| anthropic/claude-sonnet-5 | 1 | 415 | 5,804 ms | 33% |
| google/gemini-3.1-flash-lite-image | 1 | 1,829 | 2,998 ms | 33% |
On‑chain attestation
Every hour the usage roll‑up is hashed and posted to UsageAttestation on Robinhood Chain. Anyone can fetch the same roll‑up from the API, hash it, and compare — the numbers on this page are a claim you can check, not one you have to trust.
- NETWORK
- Robinhood Chain testnet · 46630
- CADENCE
- hourly epochs
- LAST EPOCH
- contract not yet deployed
- HASH
- keccak256(canonical roll‑up JSON)
Method
PCU (Parable Compute Units) estimate compute for hosted models where GPU time isn't observable: weight(tier) × (output tokens + 0.25 × input tokens). Latency is wall‑clock at the gateway including upstream time. Chain figures come from the public Robinhood Chain RPC and refresh every 15 seconds.