Qwen3.8 27B APEX needs about 13.9 GB of memory at its main file (a 11.2 GB file) with an 8K context window, so it fits on a 16 GB GPU. This repository has 2 quants, from 11.2 GB to 13.9 GB. Longer context needs more: at 32K tokens the KV cache alone adds about 8.7 GB.
Qwen3.8 27B APEX – a dense vision‑language model with 27 B parameters
Qwen3.8 27B APEX is a dense, vision‑language model based on the Qwen3.5 architecture. It contains 27 B parameters and supports up to 262 144 context tokens. The model is released under the Apache‑2.0 license and is compatible with llama.cpp and other GGUF harnesses.
Good at
- Vision‑language understanding
- Long‑horizon agentic tasks
- Coding and professional work
Limits
- No separate vision model – vision encoder included
- Requires recent llama.cpp with qwen3_5 support
- Large VRAM needed for full context
Best for
- Complex multi‑step reasoning
- Image‑and‑video analysis
- Agentic task execution
Who it’s for. Developers and researchers who need a single, dense model that can process text, images, and videos. The model is suitable for local deployment on systems with sufficient VRAM and recent llama.cpp support.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| — | 11.2 GB | 13.9 GB | Too big | Get file → | |
| — | 13.9 GB | 16.6 GB | Too big | Get file → |
The two available GGUF files are 11.24 GB (≈13.9 GB VRAM at 8 k context) and 13.94 GB (≈16.6 GB VRAM at 8 k context). For 8–12 GB VRAM, use the 11.24 GB file; for 12–18 GB VRAM, use the 13.94 GB file. Memory estimates are approximate.
What GPU do you need for Qwen3.8 27B APEX?
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 the main file (11.2 GB file, about 13.9 GB in use).
- 24 GB VRAM: the best quant that fits is the main file (13.9 GB file, about 16.6 GB in use).
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Qwen3.8 27B APEX
How much VRAM does Qwen3.8 27B APEX need?
About 13.9 GB at its main file with an 8K context window. That is the 11.2 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 Qwen3.8 27B APEX on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 13.9 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Qwen3.8 27B APEX GGUF?
It depends on the quant: file is 11.2 GB, file is 13.9 GB.
What context length does Qwen3.8 27B APEX 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 Qwen3.8 27B APEX GGUF?
The files are on Hugging Face in the mudler/Qwen3.8-27B-APEX-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 Qwen3.8 27B APEX 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 Qwen3.8 27B APEX on 8 GB VRAM?
The 11.24 GB GGUF file requires about 13.9 GB VRAM at 8 k context, so 8 GB VRAM is insufficient.
Does the model support vision tasks?
Yes, it includes a native vision encoder and can process images and videos.
What license applies to this model?
The model is released under the Apache‑2.0 license.