Subscriptions
Choose a fixed monthly budget for personal inference. Every plan includes credits that automatically stretch further when spare GPU supply is highest.
For focused personal projects and exploration.
Up to 8x your monthly payment in usage
When eligible models reach the deepest idle-supply rate.
For builders shipping and iterating regularly.
Up to 10x your monthly payment in usage
When eligible models reach the deepest idle-supply rate.
For sustained, high-volume personal workloads.
Up to 12x your monthly payment in usage
When eligible models reach the deepest idle-supply rate.
One plan per user. Non-commercial use only. Terms apply. All plans are billed monthly and can be canceled from the console.
The same inference experience, with more credits and concurrency as you grow.
| Plan feature | Basic | ProMost popular | Max |
|---|---|---|---|
| Monthly price | $10 / month | $30 / month | $100 / month |
| Included credits | $20 / month | $75 / month | $300 / month |
| Deepest idle-rate value | Up to $80 | Up to $300 | Up to $1,200 |
| Concurrent requests | 2 | 3 | 4 |
How it works
Subscription rates respond to available GPU capacity automatically. You keep building; Lilac applies the rate.
01
Your subscription renews with a fresh pool of inference credits.
02
Eligible models cost fewer credits as compatible GPUs become more idle.
03
At the deepest idle rate, your monthly credits can cover up to 4x the standard usage.
FAQ
The essentials before you choose a plan.
Personal inference
Subscribe in the Lilac console, create an API key, and use the same OpenAI-compatible API across every plan.