Skip to content

Local AI

INT8 Quantization

A quantization level that represents weights as 8-bit integers, balancing memory savings with output quality. INT8 is often used when INT4 degrades accuracy too much for a given coding task.