Cerebrium
Free tierServerless GPU infrastructure for real-time AI with sub-second cold starts and instant autoscaling
Paid·Technical·API available
Key strengths
Sub-second to low-second cold starts with GPU/memory snapshottingPay-per-second serverless GPU pricing with no idle costsElastic autoscaling across 12+ GPU types and multiple regionsNo infrastructure rewrites — bring your own code or DockerfileSOC 2, HIPAA, GDPR, ISO 27001 compliance99.999% uptime with multi-region failoverOpenTelemetry-native observability
Paid only · from $100 USD/mo
Facts verified against cerebrium.ai on
Recent changes
- Free tier: now available
- Plans updated: Pay-per-second compute pricing. Hobby plan is free (+ compute costs). Standard plan is $100/month + compute. Enterprise is custom. GPU rates range from $0.000164/s (T4) to $0.00167/s (B200). Memory at $0.00000222/GB/s, storage at $0.05/GB/mo (first 100GB free).
Get alerts when Cerebrium changes
Free. We'll email you when pricing, the free tier, API access or open-source status changes.
No ratings yet
Developer Documentation & Integration
Cerebrium exposes a full API for programmatic deployment and management of GPU workloads. Developers can deploy using existing code or a custom Dockerfile, making migration straightforward with no required infrastructure rewrites.
Supported inference frameworks & integrations:
- Inference engines: vLLM, SGLang, Triton Inference Server, TensorRT-LLM
- Real-time / voice: Twilio, Pipecat, LiveKit
- ML tooling: WandB, Gradio, FastAPI
- Observability: OpenTelemetry (native)
Infrastructure details:
- 12+ GPU types available, ranging from T4 to B200
- GPU/memory snapshotting enables sub-second to low-second cold starts
- Elastic autoscaling with multi-region failover
- 99.999% uptime SLA
Compute pricing reference:
| Resource | Rate |
|---|---|
| GPU (T4) | $0.000164/s |
| GPU (B200) | $0.00167/s |
| Memory | $0.00000222/GB/s |
| Storage | $0.05/GB/mo (first 100 GB free) |
Compliance: SOC 2, HIPAA, GDPR, ISO 27001
The project is open source with 528 GitHub stars.
Alternatives
See all alternatives