ScodyX GPU — L4 inference
$398.34/ GPU / month
A dedicated L4-class accelerator for steady inference, vision and embedding work.
- 24 GB VRAM
- Private endpoint
- Autoscale to zero
- Per-project keys
Renews at $398.34 every 1 month.
GPU & accelerators
Prices are what you pay — provisioning, renewal handling and support are included. A unit means exactly what the line says it means.
ScodyX GPU — A100 80GB
$1714.77/ GPU / month
A100 80GB capacity for fine-tuning, batch inference and heavy generation jobs.
- 80 GB VRAM
- NVLink pairs
- Fast NVMe scratch
- Reserved or on-demand
Renews at $1714.77 every 1 month.
ScodyX GPU — H100 SXM
$3511.80/ GPU / month
H100 SXM capacity for large-model training and low-latency production inference.
- 80 GB HBM3
- InfiniBand fabric
- Priority scheduling
- Reserved capacity
Renews at $3511.80 every 1 month.
ScodyX GPU Cluster
$28210.51/ node (8 GPU) / month
A managed multi-node GPU cluster with shared storage and job scheduling.
- 8 GPUs per node
- InfiniBand interconnect
- Shared fast storage
- Queue scheduler
Renews at $28210.51 every 1 month.
ScodyX GPU Serverless
$2.41/ 1,000 GPU-seconds
Per-second GPU capacity for bursty workloads, billed only while a job runs.
- Scale to zero
- Cold start under 20s
- Any container image
- Usage caps
Renews at $2.41 every 1 month.