Can starcoder2 15b i1 run on Mac mini M4 64GB?

Q: Is unified memory on Mac mini M4 64GB as fast as VRAM for starcoder2 15b i1?

Not always. Mac mini M4 64GB 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.

YES — Runs Great

C43Usable

Estimated — low-sample bucket· few comparable runs

starcoder2 15b i1 needs ~18.7 GB VRAM. Mac mini M4 64GB has 46.1 GB. With Q4_K_M quantization, expect ~9 tok/s.

Runtime: llama.cppCapacity: RoomyBandwidth: Very lowStack: StandardBottleneck: Memory bandwidth

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.

Capabilities:

Select quantization to explore

Q4_K_M (Medium quality) — 18.7 GB, 8.7 tok/s, Runs well

18.7 GB required46.1 GB available

41% VRAM used

Fit status

Runs well

Decode

8.7 tok/s

TTFT

22189 ms

Safe context

265K

Memory

18.7 GB / 46.1 GB

Memory breakdown

Weights9.2 GB

KV Cache1.8 GB

Runtime0.9 GB

Headroom6.9 GB

See how fast it feels

See how fast it feelsstarcoder2 15b i1 on Mac mini M4 64GB
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: 8.7 tok/s decode · 22.2s TTFT (warm) · 22 tok/s prefill

What limits this setup

This setup is broadly balanced for this model.

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

Performance by workload

Workload	Grade	Fit	Decode	TTFT	Context
Chat	C	Runs well	8.7 tok/s	12103 ms	265K
Coding	C	Runs well	8.7 tok/s	22189 ms	265K
Agentic Coding	C	Runs well	8.7 tok/s	32275 ms	265K
Reasoning	C	Runs well	8.7 tok/s	26224 ms	265K
RAG	C	Runs well	8.7 tok/s	40344 ms	265K

Inference speed

starcoder2 15b i1 inference speed — tokens per second by GPU & Mac

Estimated decode speed (tokens/sec) for starcoder2 15b i1 at Q4_K_M across popular GPUs and Apple Silicon, using the fastest local runtime per device. Fastest is RTX 5090 32GB at ~131 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	Q4_K_M	131.2	Fits
RTX 4090 24GB	24 GB	Q4_K_M	83.7	Fits

Quantization options

How starcoder2 15b i1 (15B params) fits at each quantization level on Mac mini M4 64GB (46.1 GB usable).

Quant	Bits	VRAM	Quality	Fit
Q2_K	2	5.9 GB	Low	C42
Q3_K_S	3	7.4 GB	Low	C42
NVFP4	4

Get started

Copy-paste commands to run starcoder2 15b i1 on your machine.

Run

lms load hf-mradermacher--starcoder2-15b-i1-gguf && lms server start

Upgrade options

Hardware that runs starcoder2 15b i1 well

MacBook Pro M4 Max 96GBBudget pick

96 GB Unified (+32)546 GB/s (+426)

Raises estimated decode speed by about 299%.34.7 tok/s decode

Raises estimated decode speed by about 299%.

Adds memory headroom for longer context windows and future model growth.

~$2,499 MSRP

MacBook Pro M3 Max 128GBBest value

128 GB Unified (+64)400 GB/s (+280)

Raises estimated decode speed by about 201%.26.2 tok/s decode

Raises estimated decode speed by about 201%.

Adds memory headroom for longer context windows and future model growth.

~$2,499 MSRP

Mac Studio M3 Ultra 96GBApple upgrade

96 GB Unified (+32)819 GB/s (+699)

Raises estimated decode speed by about 600%.60.9 tok/s decode

Raises estimated decode speed by about 600%.

Adds memory headroom for longer context windows and future model growth.

~$3,999 MSRP

Frequently asked questions

Can Mac mini M4 64GB run starcoder2 15b i1?

Yes, Mac mini M4 64GB can run starcoder2 15b i1 with a C grade (Runs well). Expected decode speed: 8.7 tok/s.

How much VRAM does starcoder2 15b i1 need?

starcoder2 15b i1 (15B parameters) requires approximately 18.7 GB of memory with Q4_K_M quantization.

What is the best quantization for starcoder2 15b i1?

The recommended quantization for starcoder2 15b i1 is Q4_K_M, which balances quality and memory efficiency.

What speed will starcoder2 15b i1 run at on Mac mini M4 64GB?

On Mac mini M4 64GB, starcoder2 15b i1 achieves approximately 8.7 tokens per second decode speed with a time-to-first-token of 22189ms using Q4_K_M quantization.

Can Mac mini M4 64GB run starcoder2 15b i1 for coding?

For coding workloads, starcoder2 15b i1 on Mac mini M4 64GB receives a C grade with 8.7 tok/s and 265K context.

What context window can starcoder2 15b i1 use on Mac mini M4 64GB?

On Mac mini M4 64GB, starcoder2 15b i1 can safely use up to 265K tokens of context. The model's official context limit is —, but available memory constrains the safe maximum.

Is unified memory on Mac mini M4 64GB as fast as VRAM for starcoder2 15b i1?

See all results for Mac mini M4 64GB See all hardware for starcoder2 15b i1

Embed this result▼

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

<iframe src="https://willitrunai.com/embed/hf-mradermacher--starcoder2-15b-i1-gguf-on-m4-mini-64gb" width="400" height="180" frameborder="0" style="border:none;border-radius:12px;overflow:hidden;" title="Will It Run AI — fit result"></iframe>

Preview: