Can Qwen3.5 397B A17B run on AMD Instinct MI325X 256GB?

YES — With NVFP4

D38Poor

Estimated from fit model

Qwen3.5 397B A17B needs ~295.3 GB VRAM. AMD Instinct MI325X 256GB has 256.0 GB. With NVFP4 quantization, expect ~12 tok/s.

Runtime: llama.cppCapacity: OffloadBandwidth: HighStack: StandardBottleneck: Host offload

Operating mode

Choose the run profile you care about

Interactive favors responsiveness, while light API and scale-out lean harder on serving readiness. The fit stays the same, but the recommendation lens changes.

Current mode

Balanced

Balanced for general local use. Keeps the ranking neutral across personal and serving workflows.

Qwen3.5 397B A17B at Q4_K_M needs 315.2 GB — too much for AMD Instinct MI325X 256GB (256.0 GB). Runs at NVFP4 (295.3 GB) with medium quality. 3 quantization levels fit.

Capabilities:

Select quantization to explore

Q4_K_M (Medium quality) — 315.2 GB, exceeds 256.0 GB available

315.2 GB required256.0 GB available

123% VRAM needed

59.2 GB over capacity — needs offload or smaller quantization

Fit status

Too heavy

Decode

8.8 tok/s

TTFT

22113 ms

Safe context

Memory

315.2 GB / 256.0 GB

Offload

20%

Memory breakdown

Weights242.2 GB

KV Cache46.5 GB

Runtime0.9 GB

Headroom25.6 GB

See how fast it feels

With memory offload — actual speed may be lower

See how fast it feelsQwen3.5 397B A17B on AMD Instinct MI325X 256GB
1st promptCold start — includes initialization
>What is local AI inference?
Local AI inference means running an AI language model directly on your own hardware — your laptop, desktop, or server — instead of sending requests to a remote cloud API.

When you run inference locally the model weights are loaded into your GPU or unified memory. Each token you generate requires reading those weights from memory, so memory bandwidth is the main bottleneck for decode speed.

Key benefits of running locally:
  - Full privacy: your prompts never leave your machine
  - No per-token cost or rate limits
  - Works offline once the model is downloaded
  - Latency depends only on your hardware
2nd promptWarm — model ready, no init overhead
>How much VRAM do I need?
It depends on the model size and quantization level. A rough rule of thumb:

  Model size     Q4 (4-bit)    Q8 (8-bit)    FP16
  7B params      ~4.3 GB       ~7.5 GB       ~14 GB
  13B params     ~7.9 GB       ~13.9 GB      ~26 GB
  70B params     ~42.7 GB      ~74.9 GB      ~140 GB

Most people use 4-bit quantization (Q4_K_M) which gives 90-95% of full quality at a fraction of the memory. A 24 GB GPU can comfortably run most 7B-13B models.
Estimated: 8.8 tok/s decode · 22.1s TTFT (warm) · 22 tok/s prefill

What limits this setup

It fits through host-memory offload, and offload is the main reason performance drops.

CPU or host-memory offload is active

About 10% of the working set spills out of accelerator memory, which usually hurts latency and sustained decode throughput.

Very little memory headroom

You can run the model, but there is not much room left for longer context, bigger batches, extra apps, or future model updates.

Best improvement path

Remove offload with more accelerator memory

Prioritize a GPU or unified-memory tier that fits the whole model natively. Removing offload usually helps more than small compute gains.

Buy headroom, not only minimum fit

A slightly larger memory tier gives you safer context growth and makes the recommendation more future-proof.

Increase host RAM if you keep offloading

This setup may need roughly 29.6 GB of extra host RAM just for the offloaded portion, before OS and other tools.

Performance by workload

Workload	Grade	Fit	Decode	TTFT	Context
Chat	D	Very compromised (needs ~29.8 GB host RAM)	10.3 tok/s	10264 ms	4K
Coding	F	Too heavy	8.8 tok/s	22113 ms	4K
Agentic Coding	F	Too heavy	6.6 tok/s	42977 ms	4K
Reasoning	F	Too heavy	8.8 tok/s	26134 ms	4K
RAG	F	Too heavy	6.6 tok/s	53722 ms	4K

Quantization options

How Qwen3.5 397B A17B (397B params) fits at each quantization level on AMD Instinct MI325X 256GB (256.0 GB usable).

Quant	Bits	VRAM	Quality	Fit
Q2_K	2	154.8 GB	Low	C48
Q3_K_SBest for your GPU	3	194.5 GB	Low	C48
NVFP4	4	222.3 GB	Medium	F0
Q4_K_M	4	242.2 GB	Medium	F0
Q5_K_M	5	285.8 GB	High	F0
Q6_K	6	325.5 GB	High	F0
Q8_0	8	424.8 GB	Very High	F0
F16	16	813.8 GB	Maximum	F0

Get started

Copy-paste commands to run Qwen3.5 397B A17B on your machine.

Run

lms load hf-unsloth--qwen3-5-397b-a17b-gguf && lms server start

アップグレードオプション

Qwen3.5 397B A17Bを快適に動かすハードウェア

AMD Instinct MI350X 288GBコスパ最良

288 GB VRAM (+32)8000 GB/s (+2000)

Makes the model fit on the accelerator instead of staying completely out of reach.14.6 tok/sデコード

Makes the model fit on the accelerator instead of staying completely out of reach.

Raises estimated decode speed by about 66%.

〜$8,000 MSRP

Frequently asked questions

Can AMD Instinct MI325X 256GB run Qwen3.5 397B A17B?

Yes, AMD Instinct MI325X 256GB can run Qwen3.5 397B A17B at NVFP4 quantization (Very compromised (needs ~29.6 GB host RAM)). The recommended Q4_K_M requires 315.2 GB which exceeds available memory, but at NVFP4 it needs only 295.3 GB. Expected decode speed: 11.5 tok/s.

How much VRAM does Qwen3.5 397B A17B need?

Qwen3.5 397B A17B (397B parameters) requires approximately 315.2 GB at Q4_K_M quantization. On AMD Instinct MI325X 256GB, it fits at NVFP4 using 295.3 GB.

What is the best quantization for Qwen3.5 397B A17B?

The recommended quantization is Q4_K_M, but on AMD Instinct MI325X 256GB the best fitting quantization is NVFP4, which uses 295.3 GB.

What speed will Qwen3.5 397B A17B run at on AMD Instinct MI325X 256GB?

On AMD Instinct MI325X 256GB, Qwen3.5 397B A17B achieves approximately 11.5 tokens per second decode speed with a time-to-first-token of 16860ms using NVFP4 quantization.

Can AMD Instinct MI325X 256GB run Qwen3.5 397B A17B for coding?

For coding workloads, Qwen3.5 397B A17B on AMD Instinct MI325X 256GB receives a F grade with 8.8 tok/s and 4K context.

What context window can Qwen3.5 397B A17B use on AMD Instinct MI325X 256GB?

On AMD Instinct MI325X 256GB, Qwen3.5 397B A17B can safely use up to 4K tokens of context at NVFP4 quantization. The model's official context limit is —, but available memory constrains the safe maximum.

What should I upgrade first if Qwen3.5 397B A17B feels slow on AMD Instinct MI325X 256GB?

Remove offload with more accelerator memory. Prioritize a GPU or unified-memory tier that fits the whole model natively. Removing offload usually helps more than small compute gains.

See all results for AMD Instinct MI325X 256GB See all hardware for Qwen3.5 397B A17B

Embed this result▼

Paste this snippet into any page to show a live fit card.

<iframe src="https://willitrunai.com/embed/hf-unsloth--qwen3-5-397b-a17b-gguf-on-instinct-mi325x-256gb" width="400" height="180" frameborder="0" style="border:none;border-radius:12px;overflow:hidden;" title="Will It Run AI — fit result"></iframe>

Preview: