GIGIlistening…

AI agents

Inference

Actually running the AI model to produce an answer.

Inference is the act of running a trained model to get an output — every reply GIGI gives is one or more inference calls under the hood. Inference is also where AI cost and latency come from, which is why metered actions on hashtag.org (long-running agent tasks, voice minutes) are billed against your Gigi credits balance.

← Back to the full glossary