Inference

The stage when a trained AI model produces an output from a new input. In production, each generated image, voice clip or video is an inference run.

In practice

Trying three versions of a prompt means running the model three times and reviewing three results.

Published: October 9, 2026 at 8:49 am
by Quint Boa, AI Video Executive & Producer