OR

Orca Cyber 27B XE2MX4

OrcasaqGGUF by cesarsal1nasUpdated 2026-10-13
Download on Hugging Face
Parameters27.3B
Context256K
Layers65
Architectureqwen35
Smallest quant16.7 GB
Quick answer

Orca Cyber 27B XE2MX4 needs about 19.3 GB of memory at its main file (a 16.7 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 8.7 GB.

Orca Cyber 27B XE2MX4 – a 27.3B Qwen35 model for Intel Xe2 GPUs

Orca Cyber 27B XE2MX4 is a 27.3B parameter Qwen35 model quantized for Intel Xe2 dGPUs. It uses a custom XE2MX4 quant format and is not compatible with standard llama.cpp, CUDA or Vulkan loaders.

Good at

  • Reasoning
  • Tools usage
  • Large context 262144 tokens

Limits

  • Requires Xe2 GPU with SYCL backend
  • Not supported by stock llama.cpp
  • Memory estimate 19.3 GB at 8 k context

Best for

  • Complex reasoning tasks
  • Tool‑integrated workflows
  • Long‑form content generation

Who it’s for. Developers building LLM applications on Intel Xe2 GPUs who need a large‑parameter model with reasoning and tool abilities. Suitable for research or production environments that can run the custom XE2MX4 loader.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
—16.7 GB 19.3 GB
Too big Get file →

The only available quant is the XE2MX4 file of size 16.66 GB. It requires approximately 19.3 GB of VRAM at 8 k context. For systems with less VRAM, this model cannot be loaded. No other quant variants are provided.

What GPU do you need for Orca Cyber 27B XE2MX4?

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 the main file (16.7 GB file, about 19.3 GB in use).

Specifications

Architecture from the GGUF header

Architectureqwen35
Parameters27.3B
Layers65
Trained context262,144 tokens
KV cache (fp16)260 KB per token

Source

Quantized bycesarsal1nas
FamilyOrcasaq
Licenseapache-2.0
Downloads (HF)689
Last updated2026-10-13

Frequently asked questions about Orca Cyber 27B XE2MX4

How much VRAM does Orca Cyber 27B XE2MX4 need?

About 19.3 GB at its main file with an 8K context window. That is the 16.7 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 Orca Cyber 27B XE2MX4 on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 19.3 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Orca Cyber 27B XE2MX4 GGUF?

It depends on the quant: file is 16.7 GB.

What context length does Orca Cyber 27B XE2MX4 support?

The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 2.2 GB at 8K and 8.7 GB at 32K.

Where can I download Orca Cyber 27B XE2MX4 GGUF?

The files are on Hugging Face in the cesarsal1nas/Orca-Cyber-27B-XE2MX4-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does Orca Cyber 27B XE2MX4 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.

Can I run Orca Cyber 27B XE2MX4 on a GPU with 8 GB VRAM?

No. The model requires about 19.3 GB of VRAM at 8 k context, which exceeds 8 GB.

What backend is needed to run this model?

It needs the SYCL backend in llama.cpp-exl3xe2cuda and the custom XE2MX4 loader from the GitHub fork.

Is the model licensed for commercial use?

Yes, it is released under the Apache‑2.0 license.

Other GGUFs of OrcaSAQ-2-Cyber-27B-Uncensored-GGUF

Similar models

Browse all →

More from cesarsal1nas

All by cesarsal1nas →

More Orcasaq models

All Orcasaq →

Other medium models

All medium models →