1792 GB/s (+992)
C
Raises estimated decode speed by about 225%.155.2 tok/s decode
Raises estimated decode speed by about 225%.
~$9,999 MSRP
Can StarCoder2 15B run on Mac Studio M2 Ultra 128GB?
YES — Runs Great
C47Usable
○Estimated from fit model
StarCoder2 15B needs ~26.7 GB VRAM. Mac Studio M2 Ultra 128GB has 92.2 GB. With Q5_K_M quantization, expect ~44 tok/s.
Runtime: llama.cppCapacity: RoomyBandwidth: HighStack: StandardBottleneck: Balanced
Share:
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.
Select quantization to explore
Q5_K_M (High quality) — 26.7 GB, 47.8 tok/s, Runs well
26.7 GB required92.2 GB available
29% VRAM used
Fit status
Runs well
Decode
47.8 tok/s
TTFT
4047 ms
Safe context
16K
Memory
26.7 GB / 92.2 GB
Memory breakdown
Weights10.8 GB
KV Cache1.2 GB
Runtime0.9 GB
Headroom13.8 GB
See how fast it feels
See how fast it feelsStarCoder2 15B on Mac Studio M2 Ultra 128GB
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: 47.8 tok/s decode · 4.0s TTFT (warm) · 120 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 | 43.8 tok/s | 2410 ms | 16K |
| Coding | C | Runs well | 43.8 tok/s | 4418 ms | 16K |
| Agentic Coding | C | Runs well | 43.8 tok/s | 6426 ms | 16K |
| Reasoning | C | Runs well | 43.8 tok/s | 5221 ms | 16K |
| RAG | C | Runs well | 43.8 tok/s | 8032 ms | 16K |
Quantization options
How StarCoder2 15B (15B params) fits at each quantization level on Mac Studio M2 Ultra 128GB (92.2 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 5.9 GB | Low | C41 |
Q3_K_S | 3 | 7.4 GB | Low | C41 |
NVFP4 | 4 |
Get started
Copy-paste commands to run StarCoder2 15B on your machine.
Run
docker run --rm -it ghcr.io/ggerganov/llama.cpp:full \
--hf-repo "bigcode/starcoder2-15b" \
--hf-file "starcoder2-15b-Q5_K_M.gguf" \
-c 4096 -ngl 99Upgrade options