Rent a GPU

Can the Apple M4 Max (64GB) run Mistral Small 3.2 24B?

Apple M4 Max (64GB) × Mistral Small 3.2 24B8K context
YesRuns fully on the GPU
Memory at Q8_027.4 GB of 48 GB
Weights 25.5 GBKV cache 1.3 GBRuntime 0.6 GB
QuantQ8_0Near-lossless
Speed~13tokens per second
Max context128Kthe 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 M4 Max (64GB) can run Mistral Small 3.2 24B fully in unified memory at Q8_0 (near-lossless), using about 27.4 GB of the 48 GB available, at roughly 13 tokens/second.

Try other context lengths

Run it with Ollama

With Ollama installed, one command downloads and starts Mistral Small 3.2 24B:

ollama run mistral-small3.2:24b

Ollama's default download is Q4_K_M (about 16.5 GB here). You have room for Q8_0, which is near-lossless: pick that GGUF in LM Studio or from Hugging Face.

New to this? Step-by-step setup for the Apple M4 Max (64GB)

Run it in the cloud

No Apple M4 Max (64GB)? Run Mistral Small 3.2 24B 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 48 GB the Apple M4 Max (64GB) can use. macOS lets the GPU use about 48 GB of the 64 GB unified memory by default.

QuantQualityMemoryFits?Speed
FP16Full quality49.9 GBToo big—
Q8_0Near-lossless27.4 GBFits~13 tok/s
Q6_KExcellent21.6 GBFits~17 tok/s
Q5_K_MVery good19 GBFits~19 tok/s
Q4_K_MGood, most popular16.5 GBFits~23 tok/s
Q3_K_MNoticeably degraded13.7 GBFits~28 tok/s
Q2_KHeavily degraded10.9 GBFits~36 tok/s

Measured speed from real Apple M4 Max (64GB) owners

Tokens per second that people measured on their own Apple M4 Max (64GB), next to our estimate.

No one has measured Mistral Small 3.2 24B on the Apple M4 Max (64GB) yet. The speeds above are estimates. Run the benchmark and yours becomes the first real number on this page.

Have an Apple M4 Max (64GB)? Add your real speed

One command times Mistral Small 3.2 24B 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 -- mistral-small3.2:24b

Windows (PowerShell)

$env:GPUFIT_MODEL="mistral-small3.2:24b"; irm https://gpufit.dev/bench.ps1 | iex

How the benchmark works

GPUs that run Mistral Small 3.2 24B comfortably

Models that run well on the Apple M4 Max (64GB)

Frequently asked questions

Can Apple M4 Max (64GB) run Mistral Small 3.2 24B?

Yes. The Apple M4 Max (64GB) can run Mistral Small 3.2 24B fully in unified memory at Q8_0 (near-lossless), using about 27.4 GB of the 48 GB available, at roughly 13 tokens/second.

How much VRAM does Mistral Small 3.2 24B need?

About 16.5 GB at Q4_K_M with 8,192 tokens of context, and about 49.9 GB at full FP16 precision.

How fast is Mistral Small 3.2 24B on the Apple M4 Max (64GB)?

Roughly 13 tokens per second at Q8_0, estimated from the Apple M4 Max (64GB)'s 546 GB/s memory bandwidth.