PR

Pristine 8B A1B MXFP4 MOE

PristineGGUF by noctrexUpdated 2025-10-27
Download on Hugging Face
Parameters8.34B
Context125K
Layers—
Architecturelfm2moe
Smallest quant0.0 GB
Quick answer

Pristine 8B A1B MXFP4 MOE needs about 5.8 GB of memory at MXFP4_MOE (a 4.8 GB file) with an 8K context window, so it fits on a 6 GB GPU. This repository has 3 quants, from 0.0 GB to 4.8 GB (MXFP4_MOE).

Pristine 8B A1B MXFP4 MOE is a 8.34B-parameter Pristine model in GGUF format, quantized by noctrex. Its trained context length is 125K tokens. The repository has 3 quant files, from 0.0 GB to 4.8 GB (MXFP4_MOE).

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
—0.0 GB 0.5 GB
Fits Get file →
MXFP4_MOE4.8 GB 5.8 GB
Fits Get file →
MXFP4_MOE4.8 GB 5.8 GB
Fits Get file →

What GPU do you need for Pristine 8B A1B MXFP4 MOE?

These figures add the file size, an fp16 KV cache at an 8K context window and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.

  • 4 GB VRAM: the best quant that fits is the main file (0.0 GB file, about 0.5 GB in use).
  • 6 GB VRAM: the best quant that fits is MXFP4_MOE (4.8 GB file, about 5.8 GB in use).

Which Pristine 8B A1B MXFP4 MOE quant should you choose?

MXFP4_MOE (4.8 GB) is the middle option in this repository. Lower bit counts mean smaller files and more quality loss.

If memory is tight, the smallest file is the first one in the table at 0.0 GB, with the quality trade-offs that come with fewer bits.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturelfm2moe
Parameters8.34B
Trained context128,000 tokens

Source

Quantized bynoctrex
FamilyPristine
Downloads (HF)93
Last updated2025-10-27

Frequently asked questions about Pristine 8B A1B MXFP4 MOE

How much VRAM does Pristine 8B A1B MXFP4 MOE need?

About 5.8 GB at MXFP4_MOE with an 8K context window. That is the 4.8 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.

Can I run Pristine 8B A1B MXFP4 MOE on an 8 GB GPU?

Yes, with MXFP4_MOE: the 4.8 GB file needs about 5.8 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of Pristine 8B A1B MXFP4 MOE GGUF?

It depends on the quant: file is 0.0 GB, MXFP4_MOE is 4.8 GB, MXFP4_MOE is 4.8 GB.

What context length does Pristine 8B A1B MXFP4 MOE support?

The GGUF header lists a trained context length of 128,000 tokens (125K).

Where can I download Pristine 8B A1B MXFP4 MOE GGUF?

The files are on Hugging Face in the noctrex/Pristine-8B-A1B-MXFP4_MOE-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

Other GGUFs of Pristine-8B-A1B

Similar models

Browse all →

More from noctrex

All by noctrex →

Other small models

All small models →