MA

Magi 36B SFT 1 DPO 1 Chkpt320

MagiGGUF by oradynUpdated 2026-03-18
Download on Hugging Face
Parameters34.7B
Context128K
Layers60
Architecturellama
Smallest quant21.0 GB
Quick answer

Magi 36B SFT 1 DPO 1 Chkpt320 needs about 23.5 GB of memory at Q4_K_M (a 21.0 GB file) with an 8K context window, so it fits on a 24 GB GPU. Longer context needs more: at 32K tokens the KV cache alone adds about 8.1 GB.

Magi 36B SFT 1 DPO 1 Chkpt320 is a 34.7B-parameter Magi model in GGUF format, quantized by oradyn. Its trained context length is 128K tokens. The repository has 1 quant file, from 21.0 GB (Q4_K_M) to 21.0 GB (Q4_K_M).

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q4_K_M21.0 GB 23.5 GB
Too big Get file →

What GPU do you need for Magi 36B SFT 1 DPO 1 Chkpt320?

These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) 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: no quant fits at an 8K context.
  • 6 GB VRAM: no quant fits at an 8K context.
  • 8 GB VRAM: no quant fits at an 8K context.
  • 12 GB VRAM: no quant fits at an 8K context.
  • 16 GB VRAM: no quant fits at an 8K context.
  • 24 GB VRAM: the best quant that fits is Q4_K_M (21.0 GB file, about 23.5 GB in use).

Which Magi 36B SFT 1 DPO 1 Chkpt320 quant should you choose?

Q4_K_M is the usual starting point for Magi 36B SFT 1 DPO 1 Chkpt320: it balances file size and quality well. Here it is a 21.0 GB file.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturellama
Parameters34.7B
Layers60
Trained context131,072 tokens
KV cache (fp16)240 KB per token

Source

Quantized byoradyn
FamilyMagi
Downloads (HF)5
Last updated2026-03-18

Frequently asked questions about Magi 36B SFT 1 DPO 1 Chkpt320

How much VRAM does Magi 36B SFT 1 DPO 1 Chkpt320 need?

About 23.5 GB at Q4_K_M with an 8K context window. That is the 21.0 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 Magi 36B SFT 1 DPO 1 Chkpt320 on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 23.5 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Magi 36B SFT 1 DPO 1 Chkpt320 GGUF?

It depends on the quant: Q4_K_M is 21.0 GB.

What context length does Magi 36B SFT 1 DPO 1 Chkpt320 support?

The GGUF header lists a trained context length of 131,072 tokens (128K). The KV cache grows with the context you use: about 2.0 GB at 8K and 8.1 GB at 32K.

Where can I download Magi 36B SFT 1 DPO 1 Chkpt320 GGUF?

The files are on Hugging Face in the oradyn/Magi-36B-SFT-1-DPO-1-chkpt320-Q4_K_M-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

Similar models

Browse all →

More from oradyn

All by oradyn →

Other large models

All large models →