Can I Run Codestral on MacBook Pro 16" (M5 Max, 128 GB)?

Superseded model. Codestral has been superseded by Devstral. This page is kept for reference; the newer family is a better starting point. View Devstral →

Written by Jakub Rusinowski · Last updated May 29, 2024

Yes — comfortably

Yes, comfortably — Codestral 22B at Q8_0 needs about 26.3 GB of the 96 GB usable on MacBook Pro 16" (M5 Max, 128 GB), leaving ~69.7 GB spare and running at ~13.2 tok/s (estimated), with room for about 32,768 tokens of context.

Confidence: high · Recommended quantization: Q8_0 · Estimated speed: ~13.2 tok/s

MacBook Pro 16" (M5 Max, 128 GB) — what it gives a model

Usable memory for models96 GB
Memory bandwidth614 GB/s
Form factorLaptop
Operating systemmacOS
Memory upgradeableNo — soldered

Codestral on MacBook Pro 16" (M5 Max, 128 GB): memory by quantization

QuantMemory neededFits 96 GB?Max contextEst. speedDownload
F1647.1 GB✓ Yes32K~7.3 tok/s44.4 GB
Q8_026.3 GB✓ Yes32K~13.2 tok/s23.6 GB
Q6_K20.9 GB✓ Yes32K~16.7 tok/s18.2 GB
Q5_K_M18.4 GB✓ Yes32K~19 tok/s15.7 GB
Q4_K_M16.1 GB✓ Yes32K~21.9 tok/s13.4 GB
Q3_K_M12.1 GB✓ Yes32K~29.5 tok/s9.5 GB
Q2_K10 GB✓ Yes32K~36.4 tok/s7.3 GB

What to watch out for

MacBook Pro M5 Max 128 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 Codestral on MacBook Pro 16" (M5 Max, 128 GB)?

Yes, comfortably — Codestral 22B at Q8_0 needs about 26.3 GB of the 96 GB usable on MacBook Pro 16" (M5 Max, 128 GB), leaving ~69.7 GB spare and running at ~13.2 tok/s (estimated), with room for about 32,768 tokens of context.

Which quantization of Codestral should I use on MacBook Pro 16" (M5 Max, 128 GB)?

Q8_0 — it needs about 26.3 GB of the 96 GB available, downloads as roughly 23.6 GB, and runs at an estimated 13.2 tokens/sec with up to 32K of context.

What limits Codestral on MacBook Pro 16" (M5 Max, 128 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 MacBook Pro 16" (M5 Max, 128 GB)

Codestral on GPUs

What This Model Is Good At

Model & Tools

← Can I Run It? | Codestral model page | Check your hardware