Hardware access
Transparent layers
Calmer WebGL backdrop, same physical pricing logic—toggle billing cadence like the home page.
HourlyMonthly
Shared Core
$0.85/ HR
Fractional access to A100/H100 clusters—ideal for bursty inference.
- Auto-scaling multi-tenant GPUs
- NVMe vector caching (100GB)
- Standard network SLA
Maximum power
Dedicated Metal
$4.20/ HR
Exclusive hardware with total isolation for peak inference.
- Bare-metal H100 SXM5 allocation
- Terabyte-scale local NVMe cache
- Secure enclave processing
Need a custom contract? Talk to solutions after you pick a layer.
Back to overview