Skip to content
Run LLM Locally

Gemma 4 26B A4B IT on the RX 7900 XT

With 20 GB of VRAM, the RX 7900 XT runs Gemma 4 26B A4B IT at Q4_K_M entirely on the GPU with an 8K context (about 18.1 GB needed). At Q8_0 it does not fit in VRAM.

ShareFacebookXWhatsAppLinkedInEmail

Gemma 4 26B A4B IT

Q4_K_M · 8K context

Compatible

The model needs about 18.1 GB and fits in the 19.4 GB available (1.3 GB left).

Your hardware
RX 7900 XT 20 GB · 64 GB RAM
Estimated memory
~18.1 GB estimate
VRAM (usable memory)
19.4 GB
GPU only
Yes
CPU/GPU offload
Not required
Generation speed
90–152 tok/s estimate
Largest context in fast memory
70K tokens
Confidence: low

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

Memory breakdown

~18.1 GB (17.4 GB–18.9 GB) needed · estimate

Fits entirely in VRAM · 1.3 GB left

  • Model weights15.9 GB
  • KV cache0.4 GB
  • Runtime overhead1 GB
  • Safety margin0.9 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
  • mixture-of-experts: speed depends on expert routing

Other quantizations on the RX 7900 XT

Gemma 4 26B A4B IT on RX 7900 XT + 64 GB RAM · 8K context
Gemma 4 26B A4B IT Q8_08K~27.7 GB · 13–25 tok/s estimateCPU/GPU offload
Gemma 4 26B A4B IT Q6_K8K~23.8 GB · 21–42 tok/s estimateCPU/GPU offload
Gemma 4 26B A4B IT Q5_K_M8K~20.4 GB · 46–94 tok/s estimateCPU/GPU offload
Gemma 4 26B A4B IT Q4_K_M8K~18.1 GB · 90–152 tok/s estimateCompatible
Gemma 4 26B A4B IT IQ4_XS8K~15.4 GB · 99–170 tok/s estimateExcellent
Gemma 4 26B A4B IT Q3_K_M8K~14.2 GB · 103–178 tok/s estimateExcellent

Alternatives

GPUs that run Gemma 4 26B A4B IT 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 Gemma 4 26B A4B 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 XT run Gemma 4 26B A4B IT?
The model needs about 18.1 GB and fits in the 19.4 GB available (1.3 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. ~18.1 GB / 19.4 GB.
What is the largest context on the RX 7900 XT?
At Q4_K_M, about 70K tokens fit entirely in VRAM with an f16 KV cache. A q8_0 KV cache roughly doubles that.
How fast is Gemma 4 26B A4B IT (RX 7900 XT)?
We estimate 90–152 tok/s (estimate, confidence low). Based on: memory bandwidth (800 GB/s) and model size.