Skip to content
Run LLM Locally

Gemma 4 31B IT on the RX 7900 XTX

With 24 GB of VRAM, the RX 7900 XTX runs Gemma 4 31B IT at Q4_K_M entirely on the GPU with an 8K context (about 20.5 GB needed). At Q8_0 it does not fit in VRAM.

ShareFacebookXWhatsAppLinkedInEmail

Gemma 4 31B IT

Q4_K_M · 8K context

Compatible

The model needs about 20.5 GB and fits in the 23.3 GB available (2.7 GB left).

Your hardware
RX 7900 XTX 24 GB · 64 GB RAM
Estimated memory
~20.5 GB estimate
VRAM (usable memory)
23.3 GB
GPU only
Yes
CPU/GPU offload
Not required
Generation speed
29–43 tok/s estimate
Largest context in fast memory
41K tokens
Confidence: low

Based on: memory bandwidth (960 GB/s) and model size

Memory breakdown

~20.5 GB (19.7 GB–21.4 GB) needed · estimate

Fits entirely in VRAM · 2.7 GB left

  • Model weights17.1 GB
  • KV cache1.4 GB
  • Runtime overhead1.1 GB
  • Safety margin1 GB
Assumptions: context 8K tokens · KV cache f16 · llama.cpp (ROCm) · flash attention on · weights = real file size · sliding-window layers keep 1,024 tokens

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 XTX

Gemma 4 31B IT on RX 7900 XTX + 64 GB RAM · 8K context
Gemma 4 31B IT Q8_08K~34.5 GB · 1.7–3 tok/s estimatePossible but slow
Gemma 4 31B IT Q6_K8K~27.3 GB · 4–8 tok/s estimateCPU/GPU offload
Gemma 4 31B IT Q5_K_M8K~23.8 GB · 14–26 tok/s estimateCPU/GPU offload
Gemma 4 31B IT Q4_K_M8K~20.5 GB · 29–43 tok/s estimateCompatible
Gemma 4 31B IT IQ4_XS8K~18.6 GB · 32–47 tok/s estimateExcellent
Gemma 4 31B IT Q3_K_M8K~17 GB · 35–51 tok/s estimateExcellent

Alternatives

GPUs that run Gemma 4 31B IT Q4_K_M entirely in VRAM

Other models that fit the RX 7900 XTX

Where to buy a RX 7900 XTX

Prices and links follow the market you choose.

Radeon RX 7900 XTX

24 GB GDDR6 · 960 GB/s · 355 W

CompatibleRuns Gemma 4 31B IT 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 XTX run Gemma 4 31B IT?
The model needs about 20.5 GB and fits in the 23.3 GB available (2.7 GB left).
Which quantization should I use on the RX 7900 XTX?
Q4_K_M is the highest-precision quantization that fits entirely in VRAM at 8K context. ~20.5 GB / 23.3 GB.
What is the largest context on the RX 7900 XTX?
At Q4_K_M, about 41K tokens fit entirely in VRAM with an f16 KV cache. A q8_0 KV cache roughly doubles that.
How fast is Gemma 4 31B IT (RX 7900 XTX)?
We estimate 29–43 tok/s (estimate, confidence low). Based on: memory bandwidth (960 GB/s) and model size.