Would CUDA be a better path than Intel Arc Pro B60 24GB for Kimi Linear 48B A3B?

Often yes, if your goal is the easiest setup and the widest runtime support. Intel can offer attractive memory capacity, but CUDA still tends to win on tooling maturity, guides, kernels, and model coverage for local AI.

Can Kimi Linear 48B A3B run on Intel Arc Pro B60 24GB?

Q: What should I upgrade first if Kimi Linear 48B A3B feels slow on Intel Arc Pro B60 24GB?

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.

YES — With Q3_K_S

B61Good

Estimated from fit model

Kimi Linear 48B A3B needs ~28.6 GB VRAM. Intel Arc Pro B60 24GB has 24.0 GB. With Q3_K_S quantization, expect ~5 tok/s.

Runtime: TransformersCapacity: OffloadBandwidth: MediumStack: 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.

Kimi Linear 48B A3B at Q4_K_M needs 34.4 GB — too much for Intel Arc Pro B60 24GB (24.0 GB). Runs at Q3_K_S (28.6 GB) with low quality. 2 quantization levels fit.

Capabilities:

Select quantization to explore

Q4_K_M (Medium quality) — 34.4 GB, exceeds 24.0 GB available

34.4 GB required24.0 GB available

143% VRAM needed

10.4 GB over capacity — needs offload or smaller quantization

Fit status

Too heavy

Decode

3.1 tok/s

TTFT

62770 ms

Safe context

Memory

34.4 GB / 24.0 GB

Offload

30%

Memory breakdown

Weights29.3 GB

KV Cache0.9 GB

Runtime1.8 GB

Headroom2.4 GB

See how fast it feels

With memory offload — actual speed may be lower

See how fast it feelsKimi Linear 48B A3B on Intel Arc Pro B60 24GB
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: 3.1 tok/s decode · 62.8s TTFT (warm) · 8 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.

Runtime ecosystem is narrower than CUDA

Intel GPUs can look attractive on memory per dollar, but local AI tooling, kernels, and model coverage are still broader and easier on CUDA today.

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.

Prefer CUDA if you want the path of least resistance

If your goal is maximum runtime coverage, easier troubleshooting, and better support for new local AI releases, CUDA is usually still the safer upgrade path.

Buy headroom, not only minimum fit

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

Performance by workload

Workload	Grade	Fit	Decode	TTFT	Context
Chat	F	Too heavy	3.2 tok/s	33304 ms	4K
Coding	F	Too heavy	3.1 tok/s	62770 ms	4K
Agentic Coding	F	Too heavy	2.9 tok/s	96390 ms	4K
Reasoning	F	Too heavy	3.1 tok/s	74183 ms	4K
RAG	F	Too heavy	2.9 tok/s	120487 ms	4K

Quantization options

How Kimi Linear 48B A3B (48B params) fits at each quantization level on Intel Arc Pro B60 24GB (24.0 GB usable).

Quant	Bits	VRAM	Quality	Fit
Q2_K	2	18.7 GB	Low	F0
Q3_K_S	3	23.5 GB	Low	F0
NVFP4	4

Get started

Copy-paste commands to run Kimi Linear 48B A3B on your machine.

Run

docker run --rm -it ghcr.io/ggerganov/llama.cpp:full \
  --hf-repo "moonshotai/Kimi-Linear-48B-A3B-Instruct" \
  --hf-file "Kimi-Linear-48B-A3B-Instruct-Q4_K_M.gguf" \
  -c 4096 -ngl 99

Upgrade options

Hardware that runs Kimi Linear 48B A3B well

Gaudi 3 128GBBudget pick

128 GB VRAM (+104)3700 GB/s (+3244)

Makes the model fit on the accelerator instead of staying completely out of reach.88.5 tok/s decode

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.

~$15,000 MSRP

Intel Data Center GPU Max 1550 128GBBest value

128 GB VRAM (+104)3200 GB/s (+2744)

Makes the model fit on the accelerator instead of staying completely out of reach.68.9 tok/s decode

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.

~$15,000 MSRP

Frequently asked questions

Can Intel Arc Pro B60 24GB run Kimi Linear 48B A3B?

Yes, Intel Arc Pro B60 24GB can run Kimi Linear 48B A3B at Q3_K_S quantization (Very compromised (needs ~3.8 GB host RAM)). The recommended Q4_K_M requires 34.4 GB which exceeds available memory, but at Q3_K_S it needs only 28.6 GB. Expected decode speed: 5.2 tok/s.

How much VRAM does Kimi Linear 48B A3B need?

Kimi Linear 48B A3B (48B parameters) requires approximately 34.4 GB at Q4_K_M quantization. On Intel Arc Pro B60 24GB, it fits at Q3_K_S using 28.6 GB.

What is the best quantization for Kimi Linear 48B A3B?

The recommended quantization is Q4_K_M, but on Intel Arc Pro B60 24GB the best fitting quantization is Q3_K_S, which uses 28.6 GB.

What speed will Kimi Linear 48B A3B run at on Intel Arc Pro B60 24GB?

On Intel Arc Pro B60 24GB, Kimi Linear 48B A3B achieves approximately 5.2 tokens per second decode speed with a time-to-first-token of 37309ms using Q3_K_S quantization.

Can Intel Arc Pro B60 24GB run Kimi Linear 48B A3B for coding?

For coding workloads, Kimi Linear 48B A3B on Intel Arc Pro B60 24GB receives a F grade with 3.1 tok/s and 4K context.

What context window can Kimi Linear 48B A3B use on Intel Arc Pro B60 24GB?

On Intel Arc Pro B60 24GB, Kimi Linear 48B A3B can safely use up to 4K tokens of context at Q3_K_S quantization. The model's official context limit is 1.0M, but available memory constrains the safe maximum.

See all results for Intel Arc Pro B60 24GB See all hardware for Kimi Linear 48B A3B

Embed this result▼

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

<iframe src="https://willitrunai.com/embed/kimi-linear-48b-a3b-on-arc-pro-b60-24gb" width="400" height="180" frameborder="0" style="border:none;border-radius:12px;overflow:hidden;" title="Will It Run AI — fit result"></iframe>

Preview: