Serverless GPU Inference Platform · Vendor file

Modal

Serverless GPU compute platform for running custom LLM fine-tuning jobs without managing servers

Active · verified June 2026

Categories Modal serves

The B4 Index places Modal in 2 categories. Each category page lays out the build-versus-buy considerations for that decision.

Serverless GPU Inference Platform Primary Per-second; H100 ~$3.95/hr; $30/mo free credit

LLM Fine-Tuning Platform GPU compute from $0.00164/GPU-second (A10G) to $0.00519/GPU-second (H100); pay-per-second with no minimums

Modal compared

Modal vs Baseten

The B4 Index scores every software category on two axes, strategic differentiation and AI feasibility, to classify it Build, Buy, Bridge, or Beware. Vendor facts here are independent of that call. See the full methodology.