Paperspace by DigitalOcean
Best for ML Notebooks, AI Model Development, Research, Computer Vision

Low-cost per-second GPU rental for training, fine-tuning, inference and AI agents
Lium is an on-demand GPU marketplace for AI developers and their agents. It runs on Bittensor Subnet 51, where independent hosts supply the GPUs. Network validators check that each GPU matches its listed specs. Renters can choose from RTX 3090, RTX 4090 and RTX 5090 up to H100, H200, B200 and B300. You can rent a single GPU or a full 8-GPU node in minutes. Lium bills per second with no minimum commitment. Prices change with supply; see live prices at lium.io/pricing. A Spot tier offers cheaper nodes that a host can reclaim. Pods run through the web app, CLI, Python SDK or API. Persistent volumes and backups keep data between pods. Nodes run in the US, Europe, South America and more. Live prices: https://lium.io/pricing?utm_source=computestacker&utm_medium=directory&utm_campaign=gpu_pricing
| GPU Model | Instance Type | Rate |
|---|---|---|
| B300 | On-Demand | $0.03/hr ~$22 /mo |
| B200 | On-Demand | $0.03/hr ~$22 /mo |
| H200 | On-Demand | $0.03/hr ~$22 /mo |
| H100 | On-Demand | $0.11/hr ~$80 /mo |
| RTX PRO 6000 | On-Demand | $0.03/hr ~$22 /mo |
| RTX 6000 Ada | On-Demand | $0.03/hr ~$22 /mo |
| L40S | On-Demand | $0.03/hr ~$22 /mo |
| RTX A6000 | On-Demand | $0.04/hr ~$29 /mo |
| RTX 5090 | On-Demand | $0.03/hr ~$22 /mo |
| RTX 4090 | On-Demand | $0.03/hr ~$22 /mo |
| RTX 3090 | On-Demand | $0.03/hr ~$22 /mo |
Rates shown above are standard on-demand pricing. This provider offers spot instances, which can reduce your costs by 30-70% if your workload is interruptible. Request a custom enterprise quote →
| Current Availability | High / Instant Provisioning |
Scores are aggregated from community benchmarks, user reviews, and technical analysis of the provider's infrastructure stack.
Sign in to ask questions, share insights, and connect with verified providers.
No discussions yet. Be the first to start the conversation!
Lium GPU pricing starts from $0.03/hr depending on GPU type, reservation model, and region.
Lium offers B300, B200, H200, H100, RTX PRO 6000, RTX 6000 Ada, L40S, RTX A6000, RTX 5090, RTX 4090, RTX 3090 GPU instances.
Lium operates in Brazil, Europe, Finland (EU), France, Global Decentralized Network, North America, South America, Sweden, United States.
Use the "Get a Quote" button to submit your requirements. You'll receive proposals within 24 hours — no commitment required.
Best for ML Notebooks, AI Model Development, Research, Computer Vision
Best for Bare Metal GPU, Low-Latency AI Inference, Global Edge AI Deployment
Best for Startups needing absolute GDPR compliance and Swiss data sovereignty.