# Live hardware status

The public hardware status page is available at [cthree.cloud/status](https://cthree.cloud/status). It shows current C3 GPU and CPU classes, the providers that can route each concrete profile, provider-specific C3 billed prices, resource ranges, high/medium/low availability tiers, and product maturity. It does not expose exact stock counts, regions, native provider SKUs, warm-pool sizes, or internal readiness reasons.

The CLI reads the same public control-plane API:

```
c3 list
c3 list --all
c3 list --class h100 --all
c3 list --class cpu --all
c3 list --provider crusoe
c3 list --provider crusoe --class l40 --all
c3 list --json
```

The default view aggregates the public provider offerings into hardware-class rows. `--all` expands them into provider/profile rows with a Provider column; `--provider` is applied by the API and composes with `--class` and `--all`. The status page uses the same provider and class filters and displays the literal equivalent CLI command.

Only configured public routes appear: the provider offering must be enabled, routing-enabled, non-spot, and snapshot-capable. A configured row can still show `LOW` when its live availability or current price is missing or stale.

Use the primary GPU class codes `l40`, `a100`, or `h100` for flexible GPU routing. A class is a routing selector; a profile is an exact C3 hardware choice within a class; a provider offering maps that profile to one provider without exposing the provider's native SKU or region. The `cpu` status class is a display group only. Use `c3 list --class cpu --all` and request one of the exact CPU hardware profile codes in `.c3`. If you submit `hardware: cpu`, C3 returns:

```
`cpu` is a hardware group, not a selectable profile.

Choose an exact CPU profile:
  c3 list --class cpu --all
```

All CPU availability is currently **experimental**. CPU profiles are usable and billable, but capacity and C3-managed images may change while the service matures. The experimental stage is separate from the live high/medium/low availability tier: an experimental CPU profile can still be currently available, and a generally available GPU can still have low live capacity.