Subscriptions

Pick your plan. Switch whenever you like.

Every plan includes GPU-hour credit at our published rates. Usage beyond the credit is billed per GPU-hour at the same price — never per media-minute.

Sandbox

Kick the tyres on the community network.

$0/month
Standard · community · ~97% uptime
from $0.20 / GPU-hour
  • API access with $5 of trial credit
  • Community supply pool only
  • 1 concurrent job
  • Email support
Standard

Batch work that can retry on drop-off.

$49/month
Standard · community · ~97% uptime
from $0.20 / GPU-hour
  • $60 of GPU-hour credit included monthly
  • Community supply pool, best-effort routing
  • 10 concurrent jobs
  • Automatic retries on node drop-off
  • Usage dashboard and per-job cost breakdown
Most popular
Institutional

Production inference with a real SLA.

$399/month
Institutional · SLA-backed · 99.9% uptime
from $0.22 / GPU-hour
  • $450 of GPU-hour credit included monthly
  • Reputation-ranked institutional hardware
  • 99.9% uptime target with auto-reroute
  • 50 concurrent jobs, priority scheduling
  • Private workload isolation and audit log
  • Shared Slack channel with our team
Scale

Reserved capacity for continuous fleets.

$1499/month
Institutional · reserved · 99.9% uptime
from Custom rate card
  • $1,800 of GPU-hour credit included monthly
  • Reserved capacity windows you schedule
  • Unlimited concurrency, dedicated queue
  • Custom workload allowlisting
  • Named solutions engineer, 1-hour response

Plans are stored on your account — no card is collected at this stage. When you're ready to start sending jobs, we'll reach out with API credentials for the tier you picked.

Account settings

Choose reliability tier

SLA-BACKED

Universities, BPOs & enterprise fleets

Reputation-ranked institutional devices with scheduled availability and auto-reroute on drop-off. Built for production inference and training jobs that can't tolerate downtime.

99.9%target uptime
Automatic matching

Tell us your workload. We'll pick the hardware.

No GPU menu to puzzle over — describe what you're running and we route it to the right hardware automatically. See exactly what was picked and why, every time.

→

Fill in your workload and click Auto-match — we'll show the exact GPU, price, and savings here.

Allowlisted workloads

A controlled job network, not a raw GPU rental.

We only run predefined, capability-matched workloads — never arbitrary code on someone else's machine. That is what makes the supply trustworthy and the pricing predictable.

Image generation

SDXL / Flux batches — GPU-saturating, tiny egress

Video generation

SVD / LTX / Wan clips priced well under datacenter spot

Embeddings

Bulk corpus backfill for search and RAG

Synthetic data

Batch LLM generation for fine-tuning shops and labs

Dataset labeling

Captioning and classification at scale

Archive backfill

Transcription, OCR, translation over huge archives

Estimate your monthly bill

Pick a GPU and how many hours a month you expect to run it. The estimate updates live with your reliability tier choice.

Estimated monthly cost
$450.00
Institutional tier · 99.9% SLA
Rate per hour$1.50/hr
Hours selected300
vs. equivalent on AWS$825.00