Skip to content
Run LLM Locally

Gemma 4 26B A4B IT on the RTX 4060

Gemma 4 26B A4B IT at Q4_K_M needs about 18.1 GB, more than the 8 GB of the RTX 4060: part of the model runs from system RAM, which slows generation. At Q3_K_M it does not fit in VRAM.

ShareFacebookXWhatsAppLinkedInEmail

Gemma 4 26B A4B IT

Q4_K_M · 8K context

Possible but slow

The model needs about 18.1 GB but only 7.7 GB of VRAM is usable: 10.4 GB must be loaded into system RAM. It runs, 42% on the GPU, but more slowly.

Your hardware
RTX 4060 8 GB · 64 GB RAM
Estimated memory
~18.1 GB estimate
VRAM (usable memory)
7.7 GB
GPU only
No
CPU/GPU offload
Required (10.4 GB)
Generation speed
11–21 tok/s estimate
Largest context in fast memory
None
Confidence: low

Based on: memory bandwidth (272 GB/s) and model size · system RAM bandwidth (~51 GB/s)

Memory breakdown

~18.1 GB (17.3 GB–18.8 GB) needed · estimate

10.4 GB must be offloaded to system RAM

  • Model weights15.9 GB
  • KV cache0.4 GB
  • Runtime overhead1 GB
  • Safety margin0.9 GB
Assumptions: context 8K tokens · KV cache f16 · llama.cpp (CUDA) · 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
  • split CPU/GPU runs vary a lot with the CPU and PCIe link
  • system RAM speed unknown (dual-channel DDR4-3200 assumed)
  • mixture-of-experts: speed depends on expert routing

Other quantizations on the RTX 4060

Gemma 4 26B A4B IT on RTX 4060 + 64 GB RAM · 8K context
Gemma 4 26B A4B IT Q8_08K~27.7 GB · 6–12 tok/s estimatePossible but slow
Gemma 4 26B A4B IT Q6_K8K~23.8 GB · 7.5–15 tok/s estimatePossible but slow
Gemma 4 26B A4B IT Q5_K_M8K~20.3 GB · 9.5–18 tok/s estimatePossible but slow
Gemma 4 26B A4B IT Q4_K_M8K~18.1 GB · 11–21 tok/s estimatePossible but slow
Gemma 4 26B A4B IT IQ4_XS8K~15.3 GB · 15–28 tok/s estimatePossible but slow
Gemma 4 26B A4B IT Q3_K_M8K~14.1 GB · 17–32 tok/s estimateCPU/GPU offload

Alternatives

GPUs that run Gemma 4 26B A4B IT Q4_K_M entirely in VRAM

Other models that fit the RTX 4060

Hardware that runs Gemma 4 26B A4B IT Q4_K_M entirely in memory

Prices and links follow the market you choose.

Radeon RX 7900 XT

AMD

CompatibleRuns Gemma 4 26B A4B IT Q4_K_M entirely in memory

Price not available

Availability unknown

View on amazon.com

Affiliate link

AMD Ryzen AI Max+ 395

Unified memory PC

ExcellentRuns 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 RTX 4060 run Gemma 4 26B A4B IT?
The model needs about 18.1 GB but only 7.7 GB of VRAM is usable: 10.4 GB must be loaded into system RAM. It runs, 42% on the GPU, but more slowly.
Which quantization should I use on the RTX 4060?
None fits entirely in VRAM at 8K context; the smallest, Q3_K_M, still needs offload.
What is the largest context on the RTX 4060?
At Q4_K_M, the model does not fit entirely in VRAM even at a short context.
How fast is Gemma 4 26B A4B IT (RTX 4060)?
We estimate 11–21 tok/s (estimate, confidence low). Based on: memory bandwidth (272 GB/s) and model size, system RAM bandwidth (~51 GB/s).