Mistral-Large-3-675B-Instruct-2512 IQ4_XS

This file is exactly 359,954,316,736 bytes — 335.23 GiB / 359.95 GB at an effective 4.276 bits per weight. The nominal rate for IQ4_XS is lower; mixed-precision tensors make the real figure higher, always.

From the file· 8 file(s), summed

Get it

335.23 GiB · 8 files
llama.cpp
llama-cli -hf unsloth/Mistral-Large-3-675B-Instruct-2512-GGUF:IQ4_XS

Downloads and runs in one step, resolving the quantization by name.

Hugging Face CLI
hf download unsloth/Mistral-Large-3-675B-Instruct-2512-GGUF --include "IQ4_XS/Mistral-Large-3-675B-Instruct-2512-IQ4_XS-*-of-*.gguf"

This quantization is split across 8 files totalling 335.23 GiB. You need every part — the first is not a smaller version of the model.

Size
335.23 GiB
359.95 GB
Effective bpw
4.276
from real bytes ÷ params
Tensors
Header
GGUF metadata