QW

chadrock 35b ace saber rocmfp4 mtp

QwenGGUF by jcbtcUpdated 2026-07-18
Download on Hugging Face
Parameters35.5B
Context256K
Layers41
Architectureqwen35moe
Smallest quant19.0 GB
Quick answer

chadrock 35b ace saber rocmfp4 mtp needs about 20.2 GB of memory at F16 (a 19.0 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 2 quants, from 19.0 GB (F16) to 31.4 GB. Longer context needs more: at 32K tokens the KV cache alone adds about 2.8 GB.

Chadrock 35B Ace Saber ROCmFP4 MTP with vision support

Chadrock 35B Ace Saber is a 35.5B‑parameter Qwen model built for AMD ROCmFP4 and ROCmFPX runners. It supports reasoning, tools, and vision when launched with the multimodal projector. The model is released under the Apache‑2.0 license.

Good at

  • Reasoning and multi‑token prediction
  • Tool‑calling
  • Vision prompts with --mmproj

Limits

  • Requires ROCmFP4/FPX runner, not stock llama.cpp
  • Large VRAM: 20.2GB at 8k context for F16
  • No native vision support without --mmproj

Best for

  • Local inference on AMD GPUs
  • Multimodal agent workflows
  • Tool‑enabled LLM applications

Who it’s for. Developers and researchers needing a large Qwen model with MTP acceleration on AMD hardware. The model is suitable for local deployment where vision and tool usage are required.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
F1619.0 GB 20.2 GB
Too big Get file →
—31.4 GB 32.6 GB
Too big Get file →

For 16GB VRAM use the F16 GGUF (19.05GB file, 20.2GB memory at 8k context). For larger memory budgets the unknown quant (31.41GB file, 32.6GB memory) provides higher precision. Both require the Chadrock v2 ROCmFPX runner.

What GPU do you need for chadrock 35b ace saber rocmfp4 mtp?

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 F16 (19.0 GB file, about 20.2 GB in use).
  • 32 GB VRAM: the best quant that fits is F16 (19.0 GB file, about 20.2 GB in use).
  • 48 GB VRAM: the best quant that fits is the main file (31.4 GB file, about 32.6 GB in use).

Which chadrock 35b ace saber rocmfp4 mtp quant should you choose?

F16 (19.0 GB) is the middle option in this repository. The unquantized model. It gives original quality and needs the most memory, so it is mainly for conversion or reference.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architectureqwen35moe
Parameters35.5B (approximate, mixture of experts)
Layers41
Trained context262,144 tokens
KV cache (fp16)82 KB per token

Source

Quantized byjcbtc
FamilyQwen
Licenseapache-2.0
Downloads (HF)142
Last updated2026-07-18

Frequently asked questions about chadrock 35b ace saber rocmfp4 mtp

How much VRAM does chadrock 35b ace saber rocmfp4 mtp need?

About 20.2 GB at F16 with an 8K context window. That is the 19.0 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 chadrock 35b ace saber rocmfp4 mtp on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 20.2 GB, so you would need a larger GPU or CPU offloading.

What is the file size of chadrock 35b ace saber rocmfp4 mtp GGUF?

It depends on the quant: F16 is 19.0 GB, file is 31.4 GB.

What context length does chadrock 35b ace saber rocmfp4 mtp 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 chadrock 35b ace saber rocmfp4 mtp GGUF?

The files are on Hugging Face in the jcbtc/chadrock-35b-ace-saber-rocmfp4-mtp repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does chadrock 35b ace saber rocmfp4 mtp 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?

No, the smallest quant requires at least 20.2GB VRAM at 8k context.

Does it support vision prompts?

Yes, launch with --mmproj to use the included vision projector.

Can it perform tool calling?

Yes, the model has detected tools ability and can be used with llama.cpp tool support.

Similar models

Browse all →

More from jcbtc

All by jcbtc →

More Qwen models

All Qwen →

Other large models

All large models →