Can InternLM 20B run on MacBook Pro M3 Pro 36GB?

YES — With Q2_K

D38Poor

Estimated from fit model

InternLM 20B needs ~30.9 GB VRAM. MacBook Pro M3 Pro 36GB has 25.9 GB. With Q2_K quantization, expect ~9 tok/s.

Runtime: llama.cppCapacity: OffloadBandwidth: Very lowStack: StandardBottleneck: Host offload

Operating mode

Choose the run profile you care about

Interactive favors responsiveness, while light API and scale-out lean harder on serving readiness. The fit stays the same, but the recommendation lens changes.

Current mode

Balanced

Balanced for general local use. Keeps the ranking neutral across personal and serving workflows.

InternLM 20B at Q5_K_M needs 37.5 GB — too much for MacBook Pro M3 Pro 36GB (25.9 GB). Runs at Q2_K (30.9 GB) with low quality.

Capabilities:

Select quantization to explore

Q5_K_M (High quality) — 37.5 GB, exceeds 25.9 GB available

37.5 GB required25.9 GB available

145% VRAM needed

11.6 GB over capacity — needs offload or smaller quantization

Fit status

Too heavy

Decode

4.7 tok/s

TTFT

40958 ms

Safe context

Memory

37.5 GB / 25.9 GB

Offload

30%

Memory breakdown

Weights14.4 GB

KV Cache18.3 GB

Runtime0.9 GB

Headroom3.9 GB

See how fast it feels

With memory offload — actual speed may be lower

See how fast it feelsInternLM 20B on MacBook Pro M3 Pro 36GB
1st promptCold start — includes initialization
>Write a Python function to sort a list
Here's a Python function that sorts a list using quicksort:

def quicksort(arr):
    if len(arr) <= 1:
        return arr
    pivot = arr[len(arr) // 2]
    left   = [x for x in arr if x < pivot]
    middle = [x for x in arr if x == pivot]
    right  = [x for x in arr if x > pivot]
    return quicksort(left) + middle + quicksort(right)

# Example usage
numbers = [3, 6, 8, 10, 1, 2, 1]
print(quicksort(numbers))  # [1, 1, 2, 3, 6, 8, 10]

The algorithm picks a pivot, partitions the list into smaller, equal, and larger elements, then recursively sorts each partition and concatenates the results.
2nd promptWarm — model ready, no init overhead
>Now add type hints and a docstring
def quicksort(arr: list[int]) -> list[int]:
    """Sort a list using the quicksort algorithm.

    Args:
        arr: List of integers to sort.

    Returns:
        A new sorted list.
    """
    if len(arr) <= 1:
        return arr
    pivot = arr[len(arr) // 2]
    left   = [x for x in arr if x < pivot]
    middle = [x for x in arr if x == pivot]
    right  = [x for x in arr if x > pivot]
    return quicksort(left) + middle + quicksort(right)

Added type hints for the input and return type, plus a Google-style docstring.
Estimated: 4.7 tok/s decode · 41.0s TTFT (warm) · 12 tok/s prefill

What limits this setup

It fits through host-memory offload, and offload is the main reason performance drops.

CPU or host-memory offload is active

About 20% of the working set spills out of accelerator memory, which usually hurts latency and sustained decode throughput.

Very little memory headroom

You can run the model, but there is not much room left for longer context, bigger batches, extra apps, or future model updates.

Shared-memory contention still exists

The OS, browser, and inference runtime all compete for the same physical memory pool, so real-world headroom is less forgiving than raw capacity suggests.

Best improvement path

Remove offload with more accelerator memory

Prioritize a GPU or unified-memory tier that fits the whole model natively. Removing offload usually helps more than small compute gains.

Buy headroom, not only minimum fit

A slightly larger memory tier gives you safer context growth and makes the recommendation more future-proof.

Increase host RAM if you keep offloading

This setup may need roughly 1.3 GB of extra host RAM just for the offloaded portion, before OS and other tools.

Performance by workload

Workload	Grade	Fit	Decode	TTFT	Context
Chat	C	Very compromised (needs ~1.2 GB host RAM)	6.7 tok/s	15828 ms	6K
Coding	F	Too heavy	4.7 tok/s	40958 ms	6K
Agentic Coding	F	Too heavy	3.5 tok/s	80680 ms	6K
Reasoning	F	Too heavy	4.7 tok/s	48405 ms	6K
RAG	F	Too heavy	3.5 tok/s	100851 ms	6K

Inference speed

InternLM 20B inference speed — tokens per second by GPU & Mac

Estimated decode speed (tokens/sec) for InternLM 20B at Q5_K_M across popular GPUs and Apple Silicon, using the fastest local runtime per device. Fastest is RTX 5090 32GB at ~45 tok/s. Speed is memory-bandwidth bound, so cards that fit the whole model in VRAM run far faster than ones that offload to system RAM.

GPU / Mac	Memory	Quant	Speed (tok/s)	Fits?
RTX 5090 32GB	32 GB	Q5_K_M	44.5	Heavy offload
Mac Studio M3 Ultra 256GB	256 GB	Q5_K_M	39.4	Fits
Mac Studio M2 Ultra 128GB	128 GB	Q5_K_M	32.9	Fits
Mac Studio M1 Ultra 128GB	128 GB	Q5_K_M	31.2	Fits
MacBook Pro M4 Max 128GB	128 GB	Q5_K_M	30.7	Fits
MacBook Pro M4 Max 64GB	64 GB	Q5_K_M	30.7	Tight
MacBook Pro M3 Max 64GB	64 GB	Q5_K_M	17.0	Tight
MacBook Pro M4 Pro 48GB	48 GB	Q5_K_M	16.1	Heavy offload
RX 7900 XTX 24GB	24 GB	Q5_K_M	15.6	Too big
MacBook Pro M1 Max 64GB	64 GB	Q5_K_M	15.6	Tight
RTX 4090 24GB	24 GB	Q5_K_M	14.0	Too big
RTX 3090 24GB	24 GB	Q5_K_M	12.9	Too big
RTX 4080 Super 16GB	16 GB	Q5_K_M	5.3	Too big
RTX 4070 12GB	12 GB	Q5_K_M	3.3	Too big
RTX 3060 12GB	12 GB	Q5_K_M	2.2	Too big
RTX 4060 8GB	8 GB	Q5_K_M	2.0	Too big

Estimates for single-stream decoding at Q5_K_M; real tokens/sec varies with prompt length, context, batch size, and runtime build. Prompt processing (prefill) is faster than the decode figures shown here.

Quantization options

How InternLM 20B (20B params) fits at each quantization level on MacBook Pro M3 Pro 36GB (25.9 GB usable).

Quant	Bits	VRAM	Quality	Fit
Q2_K	2	7.8 GB	Low	C55
Q3_K_S	3	9.8 GB	Low	B56
NVFP4	4	11.2 GB	Medium	B57
Q4_K_M	4	12.2 GB	Medium	B57
Q5_K_M	5	14.4 GB	High	B58
Q6_KBest for your GPU	6	16.4 GB	High	B58
Q8_0	8	21.4 GB	Very High	F0
F16	16	41.0 GB	Maximum	F0

Get started

Copy-paste commands to run InternLM 20B on your machine.

Run

docker run --rm -it ghcr.io/ggerganov/llama.cpp:full \
  --hf-repo "internlm/internlm2_5-20b-chat" \
  --hf-file "internlm2_5-20b-chat-Q5_K_M.gguf" \
  -c 4096 -ngl 99

アップグレードオプション

InternLM 20Bを快適に動かすハードウェア

Mac mini M4 64GBコスパ重視

64 GB Unified (+28)

Makes the model fit on the accelerator instead of staying completely out of reach.8 tok/sデコード

Makes the model fit on the accelerator instead of staying completely out of reach.

Removes host-memory offload, which is usually the single biggest latency and throughput win.

〜$1,099 MSRP

MacBook Pro M4 Pro 64GBコスパ最良

64 GB Unified (+28)273 GB/s (+123)

Makes the model fit on the accelerator instead of staying completely out of reach.19.4 tok/sデコード

Makes the model fit on the accelerator instead of staying completely out of reach.

Removes host-memory offload, which is usually the single biggest latency and throughput win.

〜$1,599 MSRP

MacBook Pro M4 Max 96GBAppleアップグレード

96 GB Unified (+60)546 GB/s (+396)

Makes the model fit on the accelerator instead of staying completely out of reach.30.7 tok/sデコード

Makes the model fit on the accelerator instead of staying completely out of reach.

Removes host-memory offload, which is usually the single biggest latency and throughput win.

〜$2,499 MSRP

Frequently asked questions

Can MacBook Pro M3 Pro 36GB run InternLM 20B?

Yes, MacBook Pro M3 Pro 36GB can run InternLM 20B at Q2_K quantization (Very compromised (needs ~1.3 GB host RAM)). The recommended Q5_K_M requires 37.5 GB which exceeds available memory, but at Q2_K it needs only 30.9 GB. Expected decode speed: 9.2 tok/s.

How much VRAM does InternLM 20B need?

InternLM 20B (20B parameters) requires approximately 37.5 GB at Q5_K_M quantization. On MacBook Pro M3 Pro 36GB, it fits at Q2_K using 30.9 GB.

What is the best quantization for InternLM 20B?

The recommended quantization is Q5_K_M, but on MacBook Pro M3 Pro 36GB the best fitting quantization is Q2_K, which uses 30.9 GB.

What speed will InternLM 20B run at on MacBook Pro M3 Pro 36GB?

On MacBook Pro M3 Pro 36GB, InternLM 20B achieves approximately 9.2 tokens per second decode speed with a time-to-first-token of 21129ms using Q2_K quantization.

Can MacBook Pro M3 Pro 36GB run InternLM 20B for coding?

For coding workloads, InternLM 20B on MacBook Pro M3 Pro 36GB receives a F grade with 4.7 tok/s and 6K context.

What context window can InternLM 20B use on MacBook Pro M3 Pro 36GB?

On MacBook Pro M3 Pro 36GB, InternLM 20B can safely use up to 8K tokens of context at Q2_K quantization. The model's official context limit is 8K, but available memory constrains the safe maximum.

What should I upgrade first if InternLM 20B feels slow on MacBook Pro M3 Pro 36GB?

Remove offload with more accelerator memory. Prioritize a GPU or unified-memory tier that fits the whole model natively. Removing offload usually helps more than small compute gains.

Is unified memory on MacBook Pro M3 Pro 36GB as fast as VRAM for InternLM 20B?

Not always. MacBook Pro M3 Pro 36GB can often fit larger models thanks to unified memory, but a discrete GPU with dedicated high-bandwidth VRAM may still decode faster once the model fits. For this combination, the important distinction is capacity versus sustained throughput.

See all results for MacBook Pro M3 Pro 36GB See all hardware for InternLM 20B

Embed this result▼

Paste this snippet into any page to show a live fit card.

<iframe src="https://willitrunai.com/embed/internlm-20b-on-m3-pro-36gb" width="400" height="180" frameborder="0" style="border:none;border-radius:12px;overflow:hidden;" title="Will It Run AI — fit result"></iframe>

Preview: