GLM / GLM-5.3-Flash
GLM-5.3-Flash FP8 quantization
Estimated until measured during engineering
Before → After
Базовый релизПроверенный чекпойнт
320.000B
320.000B
quantizationFP8
Quality retention target
Quality retention target
Производственный артефактАртефакт
Отчет о lineage и бенчмарках
Отчет о lineage и бенчмарках
Выбор GPU
Веса и KV-кэшEstimated until measured during engineering
Estimated VRAMMeasured and certified results are delivered with the project.
Runtime-пакетированиеtransformers, vllm, sglang, tokenspeed, ktransformers