Providers and models
The providers the network announces, the models they serve, and what they have been paid on chain. Figures the node measures for itself are marked as such and kept apart from chain records.
Provider scorecard
No providers are announced yet.
Providers, hardware tier and identity key from tenzro_listProviders. Standing (identity, compute bond, role stakes, validator entry) from chain state, and latency, settled calls and reputation from the node this page reads, all through tenzro_getProviderScorecard. Paid on chain is the payments each provider received in the last 7 days, from tenzro_listPayments.
Machines by hardware tier
Every machine the node holds a signed announcement for, by the hardware tier it detected, from the smallest to the largest.
The node does not report hardware tiers yet.
Machines
The node does not serve tenzro_listPeers yet.
From tenzro_listPeers. A peer's software version is known only for the node this page reads.
Inference it has served or routed
These figures come from the node's own usage accounting (tenzro_listInferenceUsage). They are real, but they are one node's records, not chain state, and are not counted in the settled-work figures.
Calls by model
Inference calls the node recorded for each model, most first.
No calls recorded.
Mean latency by model
Total recorded latency divided by calls, per model, fastest first. A mean, not a distribution: the node does not publish per-call latency.
No latency recorded.
Models served
No model is serving on this node or announced by the network.
From tenzro_listModels.