QW

MTP Qwen3.5 35B A3B

QwenGGUF by tomngdevUpdated 2026-06-28
Download on Hugging Face
Parameters1.86B
Context256K
Layers41
Architectureqwen35moe
Smallest quant1.6 GB
Quick answer

MTP Qwen3.5 35B A3B needs about 2.8 GB of memory at Q4_K_M (a 1.6 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 4 quants, from 1.6 GB (Q4_K_M) to 3.7 GB (F16). Longer context needs more: at 32K tokens the KV cache alone adds about 2.8 GB.

MTP Qwen3.5 35B A3B – a mixture‑of‑experts GGUF model

MTP Qwen3.5 35B A3B is a Qwen family causal language model with vision encoder support. It contains 35B total parameters and 3B active parameters. The model is trained with a 262144 token context length and released under Apache‑2.0.

Good at

  • Large context handling
  • Mixture‑of‑experts inference
  • Vision‑language tasks

Limits

  • Requires matching base model to run
  • No standalone usage
  • Large VRAM needed for full precision

Best for

  • High‑throughput inference on GPUs
  • Vision‑language applications
  • Research on MTP decoding

Who it’s for. Developers needing a high‑capacity, vision‑enabled language model that can be combined with a base GGUF file for speculative decoding. The model is suitable for research and production environments that can provide the necessary GPU memory and matching base model.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q4_K_M1.6 GB 2.8 GB
Fits Get file →
Q8_02.0 GB 3.2 GB
Fits Get file →
BF163.7 GB 4.9 GB
Fits Get file →
F163.7 GB 4.9 GB
Fits Get file →

For 8 GB VRAM, use the Q4_K_M quant (file size 1.62 GB, estimated 2.8 GB memory at 8 k context). For 12 GB VRAM, Q8_0 (1.99 GB, 3.2 GB memory) is suitable. Higher precision quants BF16 and F16 require 4.9 GB memory and larger VRAM.

What GPU do you need for MTP Qwen3.5 35B A3B?

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: the best quant that fits is Q8_0 (2.0 GB file, about 3.2 GB in use).
  • 6 GB VRAM: the best quant that fits is BF16 (3.7 GB file, about 4.9 GB in use).

Which MTP Qwen3.5 35B A3B quant should you choose?

Q4_K_M is the usual starting point for MTP Qwen3.5 35B A3B: it balances file size and quality well. Here it is a 1.6 GB file.

The Q8_0 file is 2.0 GB, about 1.2 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architectureqwen35moe
Parameters1.86B (approximate, mixture of experts)
Active parameters3B
Layers41
Trained context262,144 tokens
KV cache (fp16)82 KB per token

Source

Quantized bytomngdev
FamilyQwen
Licenseapache-2.0
Downloads (HF)284
Last updated2026-06-28

Frequently asked questions about MTP Qwen3.5 35B A3B

How much VRAM does MTP Qwen3.5 35B A3B need?

About 2.8 GB at Q4_K_M with an 8K context window. That is the 1.6 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 MTP Qwen3.5 35B A3B on an 8 GB GPU?

Yes, with BF16: the 3.7 GB file needs about 4.9 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of MTP Qwen3.5 35B A3B GGUF?

It depends on the quant: Q4_K_M is 1.6 GB, Q8_0 is 2.0 GB, BF16 is 3.7 GB, F16 is 3.7 GB.

What context length does MTP Qwen3.5 35B A3B support?

The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 0.7 GB at 8K and 2.8 GB at 32K.

Where can I download MTP Qwen3.5 35B A3B GGUF?

The files are on Hugging Face in the tomngdev/MTP-Qwen3.5-35B-A3B-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 MTP Qwen3.5 35B A3B 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 this model on 8 GB VRAM?

Yes, using the Q4_K_M quant will use about 2.8 GB of VRAM at 8 k context, which fits within 8 GB.

Does this model support vision tasks?

Yes, it includes a vision encoder and is designed for vision‑language tasks.

Do I need a separate base model to use this GGUF?

Yes, the GGUF files are only the MTP component and must be paired with the matching base model.

Other GGUFs of Qwen3.5-35B-A3B

Similar models

Browse all →

More from tomngdev

All by tomngdev →

More Qwen models

All Qwen →

Other tiny models

All tiny models →