Can Qwen 3.5 27B run on RX 7900 XT 20GB?

BARELY — Tight on Memory

A81Great

Estimated from fit model

Qwen 3.5 27B needs ~22.5 GB VRAM. RX 7900 XT 20GB has 20.0 GB. With Q4_K_M quantization, expect ~17 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.

Capabilities:

Select quantization to explore

Q4_K_M (Medium quality) — 22.5 GB, 18.3 tok/s, Very compromised (needs ~1.9 GB host RAM)

22.5 GB required20.0 GB available

113% VRAM needed

2.5 GB over capacity — needs offload or smaller quantization

Fit status

Very compromised (needs ~1.9 GB host RAM)

Decode

18.3 tok/s

TTFT

10552 ms

Safe context

Memory

22.5 GB / 20.0 GB

Offload

10%

Memory breakdown

Weights16.5 GB

KV Cache3.2 GB

Runtime0.9 GB

Headroom2.0 GB

See how fast it feels

See how fast it feelsQwen 3.5 27B on RX 7900 XT 20GB
1st promptCold start — includes initialization
>Write a Python function to sort a list
Here's a Python function that sorts a list using quicksort:

def quicksort(arr):
    if len(arr) <= 1:
        return arr
    pivot = arr[len(arr) // 2]
    left   = [x for x in arr if x < pivot]
    middle = [x for x in arr if x == pivot]
    right  = [x for x in arr if x > pivot]
    return quicksort(left) + middle + quicksort(right)

# Example usage
numbers = [3, 6, 8, 10, 1, 2, 1]
print(quicksort(numbers))  # [1, 1, 2, 3, 6, 8, 10]

The algorithm picks a pivot, partitions the list into smaller, equal, and larger elements, then recursively sorts each partition and concatenates the results.
2nd promptWarm — model ready, no init overhead
>Now add type hints and a docstring
def quicksort(arr: list[int]) -> list[int]:
    """Sort a list using the quicksort algorithm.

    Args:
        arr: List of integers to sort.

    Returns:
        A new sorted list.
    """
    if len(arr) <= 1:
        return arr
    pivot = arr[len(arr) // 2]
    left   = [x for x in arr if x < pivot]
    middle = [x for x in arr if x == pivot]
    right  = [x for x in arr if x > pivot]
    return quicksort(left) + middle + quicksort(right)

Added type hints for the input and return type, plus a Google-style docstring.
Estimated: 18.3 tok/s decode · 10.6s TTFT (warm) · 46 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 {ram} 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	S	Runs with offload (needs ~0.8 GB host RAM)	21.4 tok/s	4936 ms	4K
Coding	A	Very compromised	17.0 tok/s	11396 ms	4K
Agentic Coding	F	Too heavy	13.9 tok/s	20252 ms	4K
Reasoning	A	Very compromised (needs ~1.9 GB host RAM)	18.3 tok/s	12470 ms	4K
RAG	F	Too heavy	13.9 tok/s	25315 ms	4K

Quantization options

How Qwen 3.5 27B (27B params) fits at each quantization level on RX 7900 XT 20GB (20.0 GB usable).

Quant	Bits	VRAM	Quality	Fit
Q2_K	2	10.5 GB	Low	S93
Q3_K_S	3	13.2 GB	Low	S93
NVFP4Best for your GPU	4	15.1 GB	Medium	S92
Q4_K_M	4	16.5 GB	Medium	F0
Q5_K_M	5	19.4 GB	High	F0
Q6_K	6	22.1 GB	High	F0
Q8_0	8	28.9 GB	Very High	F0
F16	16	55.4 GB	Maximum	F0

Get started

Copy-paste commands to run Qwen 3.5 27B on your machine.

Run

ollama run qwen3.5:27b

Your hardware

More models your RX 7900 XT 20GB can run

Model	Params	Grade	Decode	Capabilities
Qwen3-Coder 30B A3B Instruct	30.5B	A	40.7 tok/s

Frequently asked questions

Can RX 7900 XT 20GB run Qwen 3.5 27B?

Yes, RX 7900 XT 20GB can run Qwen 3.5 27B with a A grade (Very compromised). Expected decode speed: 17.0 tok/s.

How much VRAM does Qwen 3.5 27B need?

Qwen 3.5 27B (27B parameters) requires approximately 22.5 GB of memory with Q4_K_M quantization.

What is the best quantization for Qwen 3.5 27B?

The recommended quantization for Qwen 3.5 27B is Q4_K_M, which balances quality and memory efficiency.

What speed will Qwen 3.5 27B run at on RX 7900 XT 20GB?

On RX 7900 XT 20GB, Qwen 3.5 27B achieves approximately 17.0 tokens per second decode speed with a time-to-first-token of 11396ms using Q4_K_M quantization.

Can RX 7900 XT 20GB run Qwen 3.5 27B for coding?

For coding workloads, Qwen 3.5 27B on RX 7900 XT 20GB receives a A grade with 17.0 tok/s and 4K context.

What context window can Qwen 3.5 27B use on RX 7900 XT 20GB?

On RX 7900 XT 20GB, Qwen 3.5 27B can safely use up to 4K tokens of context. The model's official context limit is 131K, but available memory constrains the safe maximum.

What should I upgrade first if Qwen 3.5 27B feels slow on RX 7900 XT 20GB?

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 RX 7900 XT 20GB See all hardware for Qwen 3.5 27B

Embed this result▼

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

<iframe src="https://willitrunai.com/embed/qwen-3.5-27b-on-rx-7900-xt-20gb" width="400" height="180" frameborder="0" style="border:none;border-radius:12px;overflow:hidden;" title="Will It Run AI — fit result"></iframe>

Preview: