P

Prime Intellect pricing

Starting price

Usage based: sandboxes from $0.04 per vCPU hour plus memory and disk, GPUs per hour by provider, training per million tokens

Partly published

Plans and prices

Prime Intellect bills usage from prepaid credits, with no plans. Sandboxes cost $0.04 per vCPU hour, $0.013 per GiB hour of memory and $0.001 per GB hour of disk. Hosted Training is priced per million tokens by model, from $0.02 input and $0.06 output and training, and GPU instances are quoted live across 50+ providers. Reserved clusters are by request.

The open source verifiers and prime-rl libraries are free to run on your own hardware

How billing works

Prepaid credits drawn by usage: GPU instances per hour by provider, sandboxes per vCPU, GiB and GB hour, Hosted Training per million tokens by model, and inference per token

GPU instances bill per hour, deducted every minute, and training bills input, output and training tokens separately. Teams can share credits and set a billing email.

What adds to the bill

Instances are deleted automatically if credits run out, so auto top up matters. Sandboxes bill for CPU, memory and disk while running.

Cost is driven almost entirely by usage: GPU hours, the providers chosen, inference volume, and the length and complexity of training and reinforcement learning runs.

Contract and support

Prime Intellect positions itself as open infrastructure to democratize training, with much of the stack open source and free to self host. It partners with clouds such as Nebius for elastic frontier hardware.

What Prime Intellect does not state

Per model inference rates and reserved cluster pricing.

GPU prices vary by provider, card and region, so there is no single headline rate.

Prime Intellect against its alternatives

Starting prices for Prime Intellect and the products closest to it, each in its own currency and billing unit.

  • Prime IntellectUsage based: sandboxes from $0.04 per vCPU hour plus memory and disk, GPUs per hour by provider, training per million tokens
  • ModalStarter is free with $30 a month of compute, Team is $250 a month, and Enterprise is custom, with usage billed per second.
  • BeamDeveloper $0 plus usage; Team $89 a month plus usage; sandboxes from $0.319 an hour
  • DaytonaUsage based, from $0.0504 per vCPU hour, with $200 of free compute
  • Together Code SandboxUsage based at $0.0446 per vCPU hour and $0.0149 per GiB of RAM an hour
  • Vercel SandboxFree allotment on Hobby

All Prime Intellect alternatives →

Contact us

Found a vendor we missed? Have feedback on the index? We'd love to hear from you.