MI

Mirxa3.6 27B

MirxaGGUF by mirxa2Updated 2026-05-04
Download on Hugging Face
Parameters26.9B
Context256K
Layers64
Architectureqwen35
Smallest quant10.0 GB
Quick answer

Mirxa3.6 27B needs about 16.9 GB of memory at Q3_K (a 14.3 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 10 quants, from 10.0 GB (IQ2_M) to 32.0 GB (Q8_K). Longer context needs more: at 32K tokens the KV cache alone adds about 8.6 GB.

Mirxa3.6 27B is a 26.9B-parameter Mirxa model in GGUF format, quantized by mirxa2. Its trained context length is 256K tokens. The repository has 10 quant files, from 10.0 GB (IQ2_M) to 32.0 GB (Q8_K).

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ2_M10.0 GB 12.7 GB
Too big Get file →
Q2_K11.5 GB 14.2 GB
Too big Get file →
IQ3_XS12.0 GB 14.6 GB
Too big Get file →
IQ3_M12.6 GB 15.2 GB
Too big Get file →
Q3_K14.3 GB 16.9 GB
Too big Get file →
IQ4_XS15.1 GB 17.7 GB
Too big Get file →
Q4_K17.5 GB 20.2 GB
Too big Get file →
Q5_K20.8 GB 23.5 GB
Too big Get file →
Q6_K23.2 GB 25.8 GB
Too big Get file →
Q8_K32.0 GB 34.6 GB
Too big Get file →

What GPU do you need for Mirxa3.6 27B?

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: the best quant that fits is IQ3_M (12.6 GB file, about 15.2 GB in use).
  • 24 GB VRAM: the best quant that fits is Q5_K (20.8 GB file, about 23.5 GB in use).
  • 32 GB VRAM: the best quant that fits is Q6_K (23.2 GB file, about 25.8 GB in use).
  • 48 GB VRAM: the best quant that fits is Q8_K (32.0 GB file, about 34.6 GB in use).

Which Mirxa3.6 27B quant should you choose?

Q3_K (14.3 GB) is the middle option in this repository. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.

If memory is tight, the smallest file is IQ2_M at 10.0 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architectureqwen35
Parameters26.9B
Layers64
Trained context262,144 tokens
KV cache (fp16)256 KB per token

Source

Quantized bymirxa2
FamilyMirxa
Licenseapache-2.0
Downloads (HF)328
Last updated2026-05-04

Frequently asked questions about Mirxa3.6 27B

How much VRAM does Mirxa3.6 27B need?

About 16.9 GB at Q3_K with an 8K context window. That is the 14.3 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 Mirxa3.6 27B on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 12.7 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Mirxa3.6 27B GGUF?

It depends on the quant: IQ2_M is 10.0 GB, Q2_K is 11.5 GB, IQ3_XS is 12.0 GB, IQ3_M is 12.6 GB, Q3_K is 14.3 GB, IQ4_XS is 15.1 GB and 4 more.

What context length does Mirxa3.6 27B support?

The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 2.1 GB at 8K and 8.6 GB at 32K.

Where can I download Mirxa3.6 27B GGUF?

The files are on Hugging Face in the mirxa2/Mirxa3.6-27B repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does Mirxa3.6 27B use?

Hugging Face lists the license as apache-2.0. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.

Similar models

Browse all →

More from mirxa2

All by mirxa2 →

More Mirxa models

All Mirxa →

Other medium models

All medium models →