selfhost.directory

The index of self-hostable projects — ranked, categorized and served with live updates.
Own your data and stop giving it to big tech.

pfnet logo

Can the Apple M4 (16 GB) run plamo-3-8b-fin-instruct?

plamo-3-8b-fin-instruct·8.5B·Apple M4 (16 GB)·16 GB unified memory

Saved in this browser, used on every model page.
Yes

Yes. The Apple M4 (16 GB) runs plamo-3-8b-fin-instruct fully in unified memory at Q8_0 (about 9.9 GB of 11 GB), at roughly 7.3 tokens/s.

best quant
Q8_0
needs
9.9 GB
usable
11 GB
est. speed
~7.3 tok/s

Estimates for an 8K context; offload assumes 32 GB of system RAM.Try other contexts on the plamo-3-8b-fin-instruct page →

Every quantisation of plamo-3-8b-fin-instruct on the Apple M4 (16 GB)

QuantQualityMemoryFitSpeed
Q2_K Heavy loss4.4 GBfits~24 tok/s
Q3_K_M Noticeable loss5.2 GBfits~16 tok/s
Q4_K_M Small loss6.2 GBfits~13 tok/s
Q5_K_M Near-lossless7.0 GBfits~11 tok/s
Q6_K Near-lossless7.8 GBfits~9.4 tok/s
Q8_0 pickLossless9.9 GBfits~7.3 tok/s
FP16 / BF16 Lossless18 GBtoo big—

Speed is a bandwidth-based estimate for generating text (120 GB/s); prompt processing and long contexts are slower.

Measured speed on the Apple M4 (16 GB)

No community reports for this pair yet.

Report your own speed →

Run plamo-3-8b-fin-instruct on the Apple M4 (16 GB)

Using Q8_0, the best fit for Apple M4 (16 GB).

Ollama · Q8_0 straight from Hugging Facedocs
ollama run hf.co/mradermacher/plamo-3-8b-fin-instruct-GGUF:Q8_0
llama.cpp · OpenAI-compatible server with a web UIdocs
llama-server -hf mradermacher/plamo-3-8b-fin-instruct-GGUF:Q8_0

brew install llama.cpp ships with Metal. Downloads the GGUF on first run.

LM Studio · Download with the lms CLIdocs
lms get mradermacher/plamo-3-8b-fin-instruct-GGUF@Q8_0

Then load it from the LM Studio app or with `lms load`.

Transformers · Chat with the pipeline APIdocs
from transformers import pipeline

pipe = pipeline("text-generation", model="pfnet/plamo-3-8b-fin-instruct", 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 (16 GB) comfortably

All that fit →
ModelParamsBest quantMemorySpeedLicence
8.2BQ8_0≈10 GBFits in memory~7.6 tok/sCommercial OK
8BQ8_0≈9.8 GBFits in memory~7.7 tok/sConditions
8BQ8_0≈9.8 GBFits in memory~7.7 tok/sConditions
7.6BQ8_0≈8.9 GBFits in memory~8.2 tok/sCommercial OK
9.4BQ6_K≈8.9 GBFits in memory~8.5 tok/sCommercial OK
9.2BQ5_K_M≈9.6 GBFits in memory~10 tok/sConditions

Other GPUs that run plamo-3-8b-fin-instruct

Frequently asked questions

Can the Apple M4 (16 GB) run plamo-3-8b-fin-instruct?
Yes. The Apple M4 (16 GB) runs plamo-3-8b-fin-instruct fully in unified memory at Q8_0 (about 9.9 GB of 11 GB), at roughly 7.3 tokens/s.
How fast does plamo-3-8b-fin-instruct run on the Apple M4 (16 GB)?
An estimate from memory bandwidth gives about 7.3 tokens/s at Q8_0.
Which quantisation of plamo-3-8b-fin-instruct should I use on the Apple M4 (16 GB)?
Q8_0: 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.