Can the Apple M4 Max 40-core (128 GB) run MLExpert?
MLExpert·34.7B (3.5B active)·Apple M4 Max 40-core (128 GB)·128 GB unified memory
Yes. The Apple M4 Max 40-core (128 GB) runs MLExpert fully in unified memory at FP16 / BF16 (about 69 GB of 96 GB), at roughly 43 tokens/s.
- best quant
- FP16 / BF16
- needs
- 69 GB
- usable
- 96 GB
- est. speed
- ~43 tok/s
Estimates for an 8K context; offload assumes 32 GB of system RAM.Try other contexts on the MLExpert page →
Every quantisation of MLExpert on the Apple M4 Max 40-core (128 GB)
| Quant | Memory | Fit | Speed |
|---|---|---|---|
| Q2_K | 15 GB | fits | ~264 tok/s |
| Q3_K_M | 18 GB | fits | ~178 tok/s |
| Q4_K_M | 23 GB | fits | ~143 tok/s |
| Q5_K_M | 28 GB | fits | ~122 tok/s |
| Q6_K | 31 GB | fits | ~106 tok/s |
| Q8_0 | 37 GB | fits | ~82 tok/s |
| FP16 / BF16 pick | 69 GB | fits | ~43 tok/s |
Speed is a bandwidth-based estimate for generating text (546 GB/s); prompt processing and long contexts are slower.
Measured speed on the Apple M4 Max 40-core (128 GB)
No community reports for this pair yet.
Report your own speed →Run MLExpert on the Apple M4 Max 40-core (128 GB)
Using FP16 / BF16, the best fit for Apple M4 Max 40-core (128 GB).
ollama run hf.co/bartowski/InternScience_MLExpert-GGUF:BF16 llama-server -hf bartowski/InternScience_MLExpert-GGUF:BF16brew install llama.cpp ships with Metal. Downloads the GGUF on first run.
lms get bartowski/InternScience_MLExpert-GGUF@BF16Then load it from the LM Studio app or with `lms load`.
from transformers import pipeline
pipe = pipeline("text-generation", model="InternScience/MLExpert", dtype="auto", device_map="auto")
messages = [{"role": "user", "content": "Explain self-hosting in one paragraph."}]
out = pipe(messages, max_new_tokens=512)
print(out[0]["generated_text"][-1]["content"])Chat models that fit the Apple M4 Max 40-core (128 GB) comfortably
All that fit →| Model | Params | Best quant | Memory | Speed |
|---|---|---|---|---|
| 36BA3B | Q8_0 | ≈38 GBFits in memory | ~94 tok/s | |
| 31.2BA4.6B | Q8_0 | ≈39 GBFits in memory | ~61 tok/s | |
| 32.8B | Q8_0 | ≈37 GBFits in memory | ~8.6 tok/s | |
| 39.8BA1.3B | Q8_0 | ≈43 GBFits in memory | ~226 tok/s | |
| 20.9BA4.2B | Q8_0 | ≈13 GBFits in memory | ~67 tok/s | |
| 30.5BA3B | Q8_0 | ≈33 GBFits in memory | ~94 tok/s |
Other GPUs that run MLExpert
Frequently asked questions
- Can the Apple M4 Max 40-core (128 GB) run MLExpert?
- Yes. The Apple M4 Max 40-core (128 GB) runs MLExpert fully in unified memory at FP16 / BF16 (about 69 GB of 96 GB), at roughly 43 tokens/s.
- How fast does MLExpert run on the Apple M4 Max 40-core (128 GB)?
- An estimate from memory bandwidth gives about 43 tokens/s at FP16 / BF16.
- Which quantisation of MLExpert should I use on the Apple M4 Max 40-core (128 GB)?
- FP16 / BF16: the largest one that fits. Expected quality: lossless compared with the full-precision weights.
Memory and speed are estimates from the model's size and the card's memory bandwidth; real numbers depend on the runtime, drivers and context length. Community figures are medians of user reports.