v2.4 — Flash attention now generally available

Run inference at production scale, not infrastructure cost

Nimbus deploys your models to a global fleet of accelerators with sub-second cold starts, automatic scaling, and billing measured in milliseconds.

Deploy in 5 minutes → Read the docs
Trusted by teams shipping at scale
HalcyonNorthwindQuantrelMeridian LabsCobaltHelix AI
Platform

Everything you need to serve models, nothing you don't

A single API from your first prototype to a billion requests a month. Bring a checkpoint, pick a region, and ship.

Sub-second cold starts

Snapshots are kept warm in regional memory pools, so the first token of a cold replica lands in under 700 ms — no pre-warming schedule to babysit.

Global routing

Requests are routed to the lowest-latency healthy region automatically. Failover between zones happens in a single request lifecycle.

Per-millisecond billing

You pay for compute time actually consumed, measured to the millisecond. Idle capacity between requests is on us, not on your invoice.

Open model catalog

One-click deployments for hundreds of open-weight models, or upload your own artifacts — weights, tokenizer, and config versioned together.

Compliant by default

SOC 2 Type II, GDPR-ready data residency, and per-tenant encryption keys. Audit logs stream to your SIEM in real time.

Observability built in

Per-request traces, token-level latency breakdowns, and drift dashboards — exported via OpenTelemetry to wherever you already look.

Developers

A familiar API, from hello-world to a million QPS

Drop-in OpenAI-compatible endpoints for chat, embeddings, and vision. Or use the native SDK with streaming, tool calling, and structured outputs.

# install
pip install nimbus-sdk

# serve any open model in one call
from nimbus import Client

client = Client(api_key="nm_live_•••")

stream = client.chat.completions.create(
    model="nimbus/nemotron-8b",
    messages=[{"role": "user", "content": "Explain KV caching."}],
    stream=True,
    region="auto",   # lowest-latency healthy region
)

for chunk in stream:
    print(chunk.choices[0].delta.content, end="")
Pricing

Pay for what you run

No idle fees, no per-seat charges. Start on the free tier and scale as traffic grows.

Hobby
$0 / forever
  • 10,000 requests / month
  • Community models only
  • 1 concurrent request
  • Shared region pool
Start free
Pro
$49 / month
  • 5M requests included
  • All open + custom models
  • 64 concurrent requests
  • Multi-region failover
  • OpenTelemetry export
Start 14-day trial
Enterprise
Custom
  • Unlimited requests
  • Dedicated accelerators
  • On-prem / VPC deployment
  • SAML SSO & audit logs
  • 99.95% uptime SLA
Talk to sales