Skip to content
Run LLM Locally

Qwen3.8 27B on the RX 7900 XT

With 20 GB of VRAM, the RX 7900 XT runs Qwen3.8 27B at Q4_K_M entirely on the GPU with an 8K context (about 18.2 GB needed). At Q8_0 it does not fit in VRAM.

ShareFacebookXWhatsAppLinkedInEmail

Qwen3.8 27B

Q4_K_M · 8K context

Compatible

The model needs about 18.2 GB and fits in the 19.4 GB available (1.2 GB left).

Your hardware
RX 7900 XT 20 GB · 64 GB RAM
Estimated memory
~18.2 GB estimate
VRAM (usable memory)
19.4 GB
GPU only
Yes
CPU/GPU offload
Not required
Generation speed
27–40 tok/s estimate
Largest context in fast memory
25K tokens
Confidence: low

Based on: memory bandwidth (800 GB/s) and model size

Memory breakdown

~18.2 GB (17.5 GB–19 GB) needed · estimate

Fits entirely in VRAM · 1.2 GB left

  • Model weights15.7 GB
  • KV cache0.6 GB
  • Runtime overhead1.1 GB
  • Safety margin0.9 GB
Assumptions: context 8K tokens · KV cache f16 · llama.cpp (ROCm) · flash attention on · weights = real file size

EstimateData updated 40 days agoSource: Run LLM Locally engine, Hugging Face config.json + model card

What makes this estimate less reliable

  • no benchmark for this hardware and model yet

Other quantizations on the RX 7900 XT

Qwen3.8 27B on RX 7900 XT + 64 GB RAM · 8K context
Qwen3.8 27B Q8_08K~30.2 GB · 1.7–3.5 tok/s estimatePossible but slow
Qwen3.8 27B Q6_K8K~23.7 GB · 4–7.5 tok/s estimateCPU/GPU offload
Qwen3.8 27B Q5_K_M8K~21.2 GB · 7.5–14 tok/s estimateCPU/GPU offload
Qwen3.8 27B Q4_K_M8K~18.2 GB · 27–40 tok/s estimateCompatible
Qwen3.8 27B IQ4_XS8K~16.9 GB · 29–43 tok/s estimateCompatible
Qwen3.8 27B Q3_K_M8K~15.4 GB · 32–47 tok/s estimateExcellent

Alternatives

GPUs that run Qwen3.8 27B Q4_K_M entirely in VRAM

Other models that fit the RX 7900 XT

Where to buy a RX 7900 XT

Prices and links follow the market you choose.

Radeon RX 7900 XT

20 GB GDDR6 · 800 GB/s · 315 W

CompatibleRuns Qwen3.8 27B Q4_K_M entirely in memory

Price not available

Availability unknown

View on amazon.com

Affiliate link

No current offer imported for this market: the link opens a marketplace search.

Frequently asked questions

Can the RX 7900 XT run Qwen3.8 27B?
The model needs about 18.2 GB and fits in the 19.4 GB available (1.2 GB left).
Which quantization should I use on the RX 7900 XT?
Q4_K_M is the highest-precision quantization that fits entirely in VRAM at 8K context. ~18.2 GB / 19.4 GB.
What is the largest context on the RX 7900 XT?
At Q4_K_M, about 25K tokens fit entirely in VRAM with an f16 KV cache. A q8_0 KV cache roughly doubles that.
How fast is Qwen3.8 27B (RX 7900 XT)?
We estimate 27–40 tok/s (estimate, confidence low). Based on: memory bandwidth (800 GB/s) and model size.