Simple, transparent pricing

Free, Start, Pro, Developer and Max. No hidden fees. Credits you don’t use don’t vanish at month end — they carry over: one month on monthly billing, two on quarterly, and the full year if you pay annually.

Credits you don't use carry into the next month, then expire. Most services wipe them at month end.

Free
Free models powered by the people's GPUs
$0/mo
No credit card needed
Start Free
  • Free tier on Llama 3.1 8B
  • 20 requests/min
  • 128K context window
  • Community support
Start
For builders and tinkerers
$20/mo
Billed monthly
  • 110M credits a month
  • Every model we serve, no tier gating
  • 60 requests/min
  • 20% discount vs pay-per-token
  • Email support
Pro
For power users and researchers
$49/mo
Billed monthly
  • 290M credits a month
  • Every model we serve, no tier gating
  • 120 requests/min
  • 28% discount vs pay-per-token
  • Priority support
Max
For scale and custom needs
$499/mo
Billed monthly
  • 2.9B credits a month
  • 20 team seats included
  • Every model we serve, no tier gating
  • 1000 requests/min
  • 40% discount vs pay-per-token
  • Dedicated support & SLAs
  • Custom contracts on request — talk to us

Plans renew automatically at the end of each billing period until you cancel. You can cancel yourself at any time from your account — you keep the period you have already paid for, and the credits already issued stay yours. Terms · Privacy

Pay as you go

No plan, no monthly commitment. Buy credits and spend them at your own pace.

per 1M tokens
Loading the current rate…

  • One credit is one token. The same rate applies to every model we serve — no tier gating.
  • Credits do not expire.
  • You are charged on the tokens a request actually consumed, settled after it completes — not on the pre-flight estimate.
Drop-in replacement. Just change the base URL. Use any OpenAI-compatible SDK. Point base_url to https://pool.linkrra.com/v1 — that's it.