Calculator

Which Mac runs which LLM, and how fast

Two numbers decide everything about running a model locally on Apple Silicon. Whether it fits in unified memory, which is arithmetic on the parameter count. And how fast it generates, which is memory bandwidth divided by the size of the weights. Pick your Mac and your model below and the calculator does both.

Yes. Llama 3.3 70B at Q4_K_M runs on M4 Pro with 64 GB.

Expect roughly 5.7 tokens/sec of generation. The weights are 42.8 GB, the 8K tokens context adds about 1 GB, and 53.2 GB of your 64 GB is spoken for once macOS is accounted for.

The two formulas

Memory. Weights in gigabytes are parameters in billions times bits per weight divided by eight. Q4_K_M averages about 4.85 bits once GGUF metadata is counted, Q8_0 about 8.5, FP16 exactly 16. Add 15% for runtime overhead, add the KV cache for your context length, then leave 3 GB for macOS. A model that fits with nothing spare will swap to SSD the moment you paste a long document, and swapping costs far more than any quantization ever saves you.

Speed. Generating one token means reading every active weight out of memory exactly once, so bandwidth divided by weight size sets the ceiling. Sitting on top of that is a fixed cost per token that does not shrink with the model: sampling, kernel launches, framework overhead. Seconds per token is the sum of the two. Fitted against the six models we measured on a base M4, that comes out at 91.5% of theoretical bandwidth plus 3.9 ms of overhead, and reproduces all six figures to within 4%.

That second term is what most calculators leave out, and it is why their numbers stop making sense at the small end. Bandwidth alone predicts 525 tokens/sec for Llama 3.2 3B on an M5 Ultra. Ours predicts 177, because once the weights are down around two gigabytes the fixed cost per token, not the memory bus, is what you are up against. Method and raw JSON for the underlying measurements are on the benchmarks page.

Where this math is wrong

Worth knowing before you spend money on the strength of a number from a web page.

Memory and speed for every model, at Q4_K_M

Minimum memory assumes an 8K context and 3 GB left for the system. Speeds are generation, in tokens per second. A dash means the model does not fit on that chip at its maximum memory configuration.

ModelWeightsMin memoryM4M4 ProM4 Max (40-core GPU)M5 Ultra
Llama 3.2 3B1.9 GB8 GB46.486129177
Mistral 7B4.4 GB16 GB22.946.879127
Qwen 2.5 7B4.6 GB16 GB21.844.876124
Llama 3.1 8B4.9 GB16 GB20.842.974120
DeepSeek R1 Distill 8B4.9 GB16 GB20.842.974120
Qwen 2.5 14B9 GB16 GB11.725.145.883
Qwen 2.5 32B19.9 GB32 GB5.41222.945.5
Llama 3.3 70B42.8 GB56 GB5.711.223.3
Qwen 2.5 72B44.1 GB56 GB5.510.922.7
Mistral Large 274.6 GB96 GB6.513.9
Llama 3.1 405B245.5 GB288 GB4.4
DeepSeek V3406.8 GB480 GB41.1

Speeds for each chip assume its largest memory configuration: M4 at 32 GB, M4 Pro at 64 GB, M4 Max (40-core GPU) at 128 GB, M5 Ultra at 512 GB.

Apple Silicon memory bandwidth, M1 to M6

The single most useful table if you are choosing hardware, because bandwidth is what you are actually buying. Figures are Apple’s own published specifications, verified 2026-09-11.

ChipBandwidthMemory optionsFound in
M168 GB/s8, 16 GBMacBook Air, 13-inch MacBook Pro, Mac mini, iMac (2020-2021)16 GB ceiling makes 14B the hard limit, and only at Q4.
M1 Pro200 GB/s16, 32 GBMacBook Pro 14/16-inch (2021)
M1 Max400 GB/s32, 64 GBMacBook Pro 14/16-inch (2021), Mac Studio (2022)Still one of the best value-per-GB/s machines on the used market.
M1 Ultra800 GB/s64, 128 GBMac Studio (2022)
M2100 GB/s8, 16, 24 GBMacBook Air, 13-inch MacBook Pro, Mac mini (2022-2023)
M2 Pro200 GB/s16, 32 GBMacBook Pro 14/16-inch, Mac mini (2023)
M2 Max400 GB/s32, 64, 96 GBMacBook Pro 14/16-inch, Mac Studio (2023)
M2 Ultra800 GB/s64, 128, 192 GBMac Studio, Mac Pro (2023)192 GB at 800 GB/s still holds 123B-class models at Q8.
M3100 GB/s8, 16, 24 GBMacBook Air, 14-inch MacBook Pro, iMac (2023-2024)
M3 Pro150 GB/s18, 36 GBMacBook Pro 14/16-inch (2023)Bandwidth dropped 25% versus M1 Pro and M2 Pro. The weakest Pro tier for inference.
M3 Max (30-core GPU)300 GB/s36, 96 GBMacBook Pro 14/16-inch (2023)The binned M3 Max. 300 GB/s against 400 on the full chip.
M3 Max (40-core GPU)400 GB/s48, 64, 128 GBMacBook Pro 14/16-inch (2023)
M3 Ultra819 GB/s96, 256, 512 GBMac Studio (2025)512 GB was the first configuration able to hold a 671B MoE in memory.
M4120 GB/s16, 24, 32 GBMacBook Air, 14-inch MacBook Pro, Mac mini, iMac (2024-2025)The one row on this page we have measured directly. See /benchmarks.
M4 Pro273 GB/s24, 48, 64 GBMacBook Pro 14/16-inch, Mac mini (2024)64 GB is the cheapest sensible way into 70B at Q4.
M4 Max (32-core GPU)410 GB/s36, 96 GBMacBook Pro 14/16-inch (2024), Mac Studio (2025)The binned M4 Max, 410 GB/s against 546 on the full chip.
M4 Max (40-core GPU)546 GB/s48, 64, 128 GBMacBook Pro 14/16-inch (2024), Mac Studio (2025)
M5153 GB/s16, 24, 32 GBMacBook Air, 14-inch MacBook Pro (2025-2026)
M5 Pro307 GB/s24, 48, 64 GBMacBook Pro 14/16-inch, Mac mini (2026)
M5 Max (base)460 GB/s36, 48, 64, 128 GBMacBook Pro 14/16-inch, Mac Studio (2026)Configurable to 614 GB/s. Check which one you are actually buying.
M5 Max (614 GB/s)614 GB/s48, 64, 128 GBMacBook Pro 14/16-inch, Mac Studio (2026)
M5 Ultra1200 GB/s96, 256, 512 GBMac Studio (2026)1.2 TB/s and 512 GB. The fastest single box for local inference Apple has shipped.
M6153 GB/s16, 24, 32 GBMac mini (2026)Up to 170 GB/s on higher configurations; 153 is the entry figure used here.

Common questions

How much RAM do I need to run a 70B model on a Mac?

About 56 GB at Q4_K_M with a short context: roughly 43 GB of weights, 15% runtime overhead, a gigabyte or two of KV cache, and 3 GB left for macOS. That makes 64 GB the smallest practical configuration, which in the current lineup means an M4 Pro, M5 Pro, or any Max or Ultra chip. At Q8 the same model needs around 128 GB.

Does a faster GPU make local LLMs faster on a Mac?

Barely, for generation. Producing one token means reading every active weight out of memory once, so generation speed is set by memory bandwidth, not GPU cores. A 24 GB M4 with more GPU cores runs an 8B model at the same speed as a 16 GB M4 because both sit on the same 120 GB/s bus. Prompt processing is the exception: that stage is compute bound and does scale with cores.

How accurate are these speed estimates?

They reproduce our own measurements to within 4%. On a base M4 the model predicts 46.4 tokens/sec for Llama 3.2 3B against 46.7 measured, 20.8 for Llama 3.1 8B against 21.2, and 11.7 for Qwen 2.5 14B against 11.7. Those six models on that one chip are everything we have measured directly, so every other row is extrapolation from the same two constants rather than a reading off a machine. Treat them as planning numbers, accurate enough to choose hardware by and not precise enough to quote.

Why does the calculator reserve 3 GB for the system?

Because a model that fits with nothing to spare is a model that swaps. macOS, a browser, and whatever agent process is calling the model all need memory, and once the machine starts paging to SSD the generation speed collapses regardless of what the bandwidth math says.

Do mixture-of-experts models follow the same math?

Memory yes, speed no. An MoE model has to hold every expert in unified memory, so the total parameter count decides whether it fits. Per token it reads only the active experts, so generation speed follows the active parameter count. That is why a 671B MoE with 37B active can generate faster than a dense 70B while needing eight times the memory.

When the answer is a Mac you do not own

If the model you want needs 64 GB or more and the machine has to be reachable around the clock, buying is not the only option. We rent dedicated Apple Silicon with Ollama already installed and an OpenAI compatible endpoint already exposed, from a base M4 Mac mini at $79/mo up to a 256 GB Mac Studio. Same hardware, same math as this page, none of the hosting.