ModelRigSYS: READY
Home / Models / Mixtral 8x7B
LLM / MODEL FILE

Mixtral 8x7B

Mistral model for mixture-of-experts inference. Memory figures are estimates, not measured benchmarks.

Parameters46.7B
Q4 estimate34.1 GB
WorkloadLLM
Best proofTest exact runtime
QUANTIZATION

Memory scenarios

CLOSEST FITS

Hardware near the Q4 line

LIMITFitting in memory does not guarantee acceptable speed.

Runtime, drivers, memory bandwidth, context, batch size, offloading, and model architecture all matter.