Models live
—
public HF models on the network
Active providers
—
idle Macs earning per token
Priced from
—/M in
cheapest model · per million input tokens
Console accounts
—
accounts active on this console · not the network total
Inference requests
—
completions served · metered on the coordinator
Tokens served
—
input + output · metered on the coordinator
Console API keys
—
active keys on this console · not the network total
Hardware attestation verify →
The coordinator runs inside a confidential VM — proven live below, re-checkable in your own browser. Each provider response additionally carries its own hardware-rooted receipt once providers are live; both are checkable with no trust in Umbra required.
checking…in_cvmmodel digest-verified
reading live proof from /v1/attestation …
Price spread $/M input tokens · live catalog
Each bar is one model's input price per million tokens — the long tail, priced low.
—avg ——
Model catalog full catalog →
Live from the coordinator registry. Public HF repos, pinned to a known-good revision and GGUF digest, priced per million tokens.
| Model | Arch | Price in | Price out | Status |
|---|---|---|---|---|
| — | — | — | — | — |
| — | — | — | — | — |
| — | — | — | — | — |