Can I Run Gemma 2 Family on Mac Studio (M3 Ultra, 256 GB)?

Superseded model. Gemma 2 Family has been superseded by Gemma 4. This page is kept for reference; the newer family is a better starting point. View Gemma 4 →

Written by Jakub Rusinowski · Last updated June 27, 2024

Yes — comfortably

Yes, comfortably — Gemma 2 9B IT at Q8_0 needs about 13.2 GB of the 192 GB usable on Mac Studio (M3 Ultra, 256 GB), leaving ~178.8 GB spare and running at ~36.4 tok/s (estimated), with room for about 8,192 tokens of context.

Confidence: high · Recommended quantization: Q8_0 · Estimated speed: ~36.4 tok/s

Mac Studio (M3 Ultra, 256 GB) — what it gives a model

Usable memory for models192 GB
Memory bandwidth819 GB/s
Form factorDesktop
Operating systemmacOS
Memory upgradeableNo — soldered

Gemma 2 Family on Mac Studio (M3 Ultra, 256 GB): memory by quantization

QuantMemory neededFits 192 GB?Max contextEst. speedDownload
F1621.6 GB✓ Yes8K~21.6 tok/s18 GB
Q8_013.2 GB✓ Yes8K~36.4 tok/s9.6 GB
Q6_K11 GB✓ Yes8K~44.3 tok/s7.4 GB
Q5_K_M10 GB✓ Yes8K~49.1 tok/s6.4 GB
Q4_K_M9.1 GB✓ Yes8K~54.8 tok/s5.4 GB
Q3_K_M7.5 GB✓ Yes8K~68.1 tok/s3.8 GB
Q2_K6.6 GB✓ Yes8K~78.6 tok/s3 GB

What to watch out for

Mac Studio M3 Ultra 256 GB limitations

Recommended setup

Ollama or LM Studio (Metal) — MLX for the fastest Apple-native throughput

How these numbers are calculated

FAQ

Can I run Gemma 2 Family on Mac Studio (M3 Ultra, 256 GB)?

Yes, comfortably — Gemma 2 9B IT at Q8_0 needs about 13.2 GB of the 192 GB usable on Mac Studio (M3 Ultra, 256 GB), leaving ~178.8 GB spare and running at ~36.4 tok/s (estimated), with room for about 8,192 tokens of context.

Which quantization of Gemma 2 Family should I use on Mac Studio (M3 Ultra, 256 GB)?

Q8_0 — it needs about 13.2 GB of the 192 GB available, downloads as roughly 9.6 GB, and runs at an estimated 36.4 tokens/sec with up to 8K of context.

What limits Gemma 2 Family on Mac Studio (M3 Ultra, 256 GB)?

Nothing binding — the model fits with headroom and generates at a usable speed on this hardware.

Which runtime should I use?

Ollama or LM Studio (Metal) — MLX for the fastest Apple-native throughput

Other Computers

Other Models on Mac Studio (M3 Ultra, 256 GB)

Gemma 2 Family on GPUs

What This Model Is Good At

Model & Tools

← Can I Run It? | Gemma 2 Family model page | Check your hardware