GPU & inference hosting

Rent GPUs to run local models with Ollama or vLLM behind your agent.

Price sorting is approximate: monthly plans are ordered by their entry price (EUR shown ×1.1 for ordering only). Usage-based sandboxes and GPU hosts have no monthly price and sort last. Always check the provider's page.

• Lambda

  Straightforward GPU rental for serious local inference.
  from $0.69/GPU-hr (RTX 6000 24 GB)
  GPU Price verified 2026-10-16

  GPUUsage-based
  Details

• RunPod

  Cheap GPUs for running Ollama or vLLM for your agent.
  from $0.27/hr (RTX A5000 24 GB)
  GPU Price verified 2026-10-16

  GPUOllamaUsage-based
  Details

• Vast.ai

  Cheapest GPU hours if you can tolerate interruption.
  Live 'from' rates, e.g. RTX 5090 from $0.39/hr; H100 SXM from $1.73/hr
  GPU Price verified 2026-10-16

  GPUMarketplaceUsage-based
  Details