Signal & Noise
Menu

Batch inference

Processing many requests together without a latency requirement, typically at ~50% discount on provider batch APIs. The right home for any AI workload a user isn't actively waiting on.

Related terms

← Back to the full glossary