The best tools for AI developers
Serverless RL runs the rollout and training phases on separate infrastructure, eliminating GPU idle time caused by lingering rollouts. You pay only for active usage, not idle time
GPU-hours are calculated by aggregating the total time used to train your models during the last billing cycle. Training a single step requires GPU time for three actions: downloading the most recent LoRA to train from, adjusting the LoRA weights using GRPO, and saving the updated weights. Since the downloading and saving processes only take a few seconds each, the bulk of a training step is dedicated to actually training your model.
No, jobs are billed for the GPU time they use, with no minimum training duration.
GPU time for failed jobs will not be charged to the user’s account.
A token is a mathematical representation of natural language. Log in to your account to view your billing dashboard. This dashboard will show you how many tokens you’ve used during the current and past months.
