M5 generation
Local LLMs on the M5 Max (614 GB/s)
The M5 Max (614 GB/s) moves 614 GB/s and takes up to 128 GB of unified memory. That combination runs everything up to Mistral Large 2 at Q4, at roughly 7.3 tokens/sec. Bandwidth sets the speed, memory sets the ceiling, and the tables below give both for every model.
Memory bandwidth
614 GB/s
Unified memory
48, 64, 128 GB
8B model at Q4
~80 tok/s
Found in: MacBook Pro 14/16-inch, Mac Studio (2026). Specifications verified 2026-09-11against Apple’s published figures.
What the M5 Max (614 GB/s) runs, at Q4_K_M
Memory figures include a 3 GB reserve for macOS and an 8K context. Speeds come from a model fitted to our own measured benchmarks; the method and its error bars are on the calculator page.
| Model | Needs | 48 GB | 64 GB | 128 GB |
|---|---|---|---|---|
| Llama 3.2 3B | 8 GB | 136 tok/s | 136 tok/s | 136 tok/s |
| Mistral 7B | 16 GB | 86 tok/s | 86 tok/s | 86 tok/s |
| Qwen 2.5 7B | 16 GB | 83 tok/s | 83 tok/s | 83 tok/s |
| Llama 3.1 8B | 16 GB | 80 tok/s | 80 tok/s | 80 tok/s |
| DeepSeek R1 Distill 8B | 16 GB | 80 tok/s | 80 tok/s | 80 tok/s |
| Qwen 2.5 14B | 16 GB | 50 tok/s | 50 tok/s | 50 tok/s |
| Qwen 2.5 32B | 32 GB | 25.5 tok/s | 25.5 tok/s | 25.5 tok/s |
| Llama 3.3 70B | 56 GB | — | 12.5 tok/s | 12.5 tok/s |
| Qwen 2.5 72B | 56 GB | — | 12.1 tok/s | 12.1 tok/s |
| Mistral Large 2 | 96 GB | — | — | 7.3 tok/s |
| Llama 3.1 405B | 288 GB | — | — | — |
| DeepSeek V3 | 480 GB | — | — | — |
What it cannot run
Even at 128 GB, these need more unified memory than the M5 Max (614 GB/s) can address at Q4: Llama 3.1 405B, DeepSeek V3. Dropping the quantization buys a little room but not a generation of it, so the real options are a chip with a higher ceiling or several machines with pooled memory.
M5 Max (614 GB/s) against the M4 Max (40-core GPU)
Bandwidth went from 546 GB/s to 614 GB/s, which is 1.12x, and that ratio carries almost directly into generation speed at the same model and quantization. The memory ceiling went from 128 GB to 128 GB, which is the part that decides whether a model runs at all rather than how fast. The ceiling did not move, so this is a speed upgrade rather than a capability one. If a model did not fit before, it still will not.
Common questions
What is the largest model an M5 Max (614 GB/s) can run?
At Q4_K_M with 128 GB of unified memory, Mistral Large 2 is the largest of the major open models that fits, needing about 96 GB once macOS is accounted for. Anything larger has to drop to a smaller quantization, run across clustered machines, or move to a chip with a higher memory ceiling.
How fast is the M5 Max (614 GB/s) for local inference?
It has 614 GB/s of memory bandwidth, and generation speed on Apple Silicon is set almost entirely by bandwidth divided by the size of the weights. In practice that works out to roughly 80 tokens/sec on an 8B model at Q4 and 25.5 on a 32B, assuming they fit. That is about 1.12x the M4 Max (40-core GPU) at the same model size.
How much memory should I order with an M5 Max (614 GB/s)?
Memory decides which models you can run at all, and it cannot be upgraded later, so it is the one specification worth overbuying. M5 Max (614 GB/s) ships in 48, 64, 128 GB. The 48 GB configuration caps you at the small end; 128 GB is what opens up the larger models in the table on this page.
The M5 Max (614 GB/s) is in our lineup, not yet in our racks
Straight answer, since you are probably here deciding what to buy. We do not own this hardware yet. It is listed in our configurator at its real price, $357/mo for the entry configuration, and choosing it puts you on the waitlist rather than through checkout. If you want one, that is the honest way to tell us, and it is what decides which machines we buy next. Available today, the closest thing we run is the M4 Max (40-core GPU) at 546 GB/s, up to 128 GB, from $286/mo, which is 89% of the bandwidth in the table above.