Inference
Simple
A trained AI at work: the act of producing an answer.
Business
The operational phase of a model, where all request costs live.
Tech
The model’s forward pass in production; metrics: latency, throughput, cost per token.
A trained AI at work: the act of producing an answer.
The operational phase of a model, where all request costs live.
The model’s forward pass in production; metrics: latency, throughput, cost per token.