Rent a GPU

Can the Apple M3 Ultra (256GB) run Gemma 4 26B-A4B?

Apple M3 Ultra (256GB) × Gemma 4 26B-A4B8K context
YesRuns fully on the GPU
Memory at FP1652.5 GB of 192 GB
Weights 51.6 GBKV cache 0.3 GBRuntime 0.6 GB
QuantFP16Full quality
Speed~65tokens per second
Max context256Kthe model's full window
  1. FP16
  2. Q8_0
  3. Q6_K
  4. Q5_K_M
  5. Q4_K_M
  6. Q3_K_M
  7. Q2_K

Yes. The Apple M3 Ultra (256GB) can run Gemma 4 26B-A4B fully in unified memory at FP16 (full quality), using about 52.5 GB of the 192 GB available, at roughly 65 tokens/second.

Try other context lengths

Run it with Ollama

With Ollama installed, one command downloads and starts Gemma 4 26B-A4B:

ollama run gemma4:26b

Ollama's default download is Q4_K_M (about 16.5 GB here). You have room for FP16, which is full quality: pick that GGUF in LM Studio or from Hugging Face.

New to this? Step-by-step setup for the Apple M3 Ultra (256GB)

Run it in the cloud

No Apple M3 Ultra (256GB)? Run Gemma 4 26B-A4B in the cloud.

Rent a cloud GPU by the hour. No upfront cost, stop any time.

Rent a GPU on RunPodRent a GPU on Vast.ai

Memory needed by quantization

At 8,192 tokens of context. The black line marks the 192 GB the Apple M3 Ultra (256GB) can use. macOS lets the GPU use about 192 GB of the 256 GB unified memory by default.

QuantQualityMemoryFits?Speed
FP16Full quality52.5 GBFits~65 tok/s
Q8_0Near-lossless28.3 GBFits~122 tok/s
Q6_KExcellent22 GBFits~158 tok/s
Q5_K_MVery good19.2 GBFits~182 tok/s
Q4_K_MGood, most popular16.5 GBFits~213 tok/s
Q3_K_MNoticeably degraded13.5 GBFits~265 tok/s
Q2_KHeavily degraded10.6 GBFits~345 tok/s

Measured speed from real Apple M3 Ultra (256GB) owners

Tokens per second that people measured on their own Apple M3 Ultra (256GB), next to our estimate.

No one has measured Gemma 4 26B-A4B on the Apple M3 Ultra (256GB) yet. The speeds above are estimates. Run the benchmark and yours becomes the first real number on this page.

Have an Apple M3 Ultra (256GB)? Add your real speed

One command times Gemma 4 26B-A4B in your own Ollama, shows you exactly what it would send, and asks before sending anything. No sign-up, no personal data.

macOS and Linux

curl -fsSL https://gpufit.dev/bench.sh | sh -s -- gemma4:26b

Windows (PowerShell)

$env:GPUFIT_MODEL="gemma4:26b"; irm https://gpufit.dev/bench.ps1 | iex

How the benchmark works

GPUs that run Gemma 4 26B-A4B comfortably

Models that run well on the Apple M3 Ultra (256GB)

Frequently asked questions

Can Apple M3 Ultra (256GB) run Gemma 4 26B-A4B?

Yes. The Apple M3 Ultra (256GB) can run Gemma 4 26B-A4B fully in unified memory at FP16 (full quality), using about 52.5 GB of the 192 GB available, at roughly 65 tokens/second.

How much VRAM does Gemma 4 26B-A4B need?

About 16.5 GB at Q4_K_M with 8,192 tokens of context, and about 52.5 GB at full FP16 precision.

How fast is Gemma 4 26B-A4B on the Apple M3 Ultra (256GB)?

Roughly 65 tokens per second at FP16, estimated from the Apple M3 Ultra (256GB)'s 819 GB/s memory bandwidth.