Can Leanstral 119B A6B run on NVIDIA DGX Spark 128GB?

YES — Tight Fit

A81Great

Estimated from fit model

Leanstral 119B A6B needs ~96.8 GB VRAM. NVIDIA DGX Spark 128GB has 108.8 GB. With Q4_K_M quantization, expect ~5 tok/s.

Runtime: vLLMCapacity: TightBandwidth: LowStack: OptimizedBottleneck: 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) — 96.8 GB, 5.0 tok/s, Tight fit

96.8 GB required108.8 GB available

89% VRAM used

Fit status

Tight fit

Decode

5.0 tok/s

TTFT

38800 ms

Safe context

38K

Memory

96.8 GB / 108.8 GB

Memory breakdown

Weights72.6 GB

KV Cache8.8 GB

Runtime2.4 GB

Headroom13.1 GB

See how fast it feels

See how fast it feelsLeanstral 119B A6B on NVIDIA DGX Spark 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: 5.0 tok/s decode · 38.8s TTFT (warm) · 13 tok/s prefill

What limits this setup

The model fits in shared memory, but shared-memory bandwidth is now the real limiter.

Fit does not mean dedicated-VRAM speed

Unified or shared memory can make a model technically fit, but sustained tokens per second may still trail a discrete high-bandwidth GPU with less total memory.

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

Prioritize bandwidth, not only capacity

If this workload feels slow, the next useful step is often a GPU tier with materially faster memory bandwidth rather than only a small bump in capacity.

Performance by workload

Workload	Grade	Fit	Decode	TTFT	Context
Chat	A	Tight fit	5.0 tok/s	21164 ms	38K
Coding	A	Tight fit	5.0 tok/s	38800 ms	38K
Agentic Coding	F	Too heavy	5.0 tok/s	56436 ms	38K
Reasoning	A	Tight fit	5.0 tok/s	45854 ms	38K
RAG	F	Too heavy	5.0 tok/s	70545 ms	38K

Inference speed

Leanstral 119B A6B inference speed — tokens per second by GPU & Mac

Estimated decode speed (tokens/sec) for Leanstral 119B A6B at Q4_K_M across popular GPUs and Apple Silicon, including multi-GPU rigs, using the fastest local runtime per device. Fastest is Mac Studio M3 Ultra 256GB at ~17 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?
Mac Studio M3 Ultra 256GB	256 GB	Q4_K_M	17.0	Fits
Mac Studio M2 Ultra 128GB	128 GB	Q4_K_M	11.9	Too big
Mac Studio M1 Ultra 128GB	128 GB	Q4_K_M	11.3	Too big
MacBook Pro M4 Max 128GB	128 GB	Q4_K_M	8.8	Too big
2× RX 7900 XTX 24GB	48 GB	Q4_K_M	7.4	Too big
2× RTX 4090 24GB	48 GB	Q4_K_M	6.7	Too big
2× RTX 3090 24GB	48 GB	Q4_K_M	5.8	Too big
RTX 5090 32GB	32 GB	Q4_K_M	5.5	Too big
4× RTX 3060 12GB	48 GB	Q4_K_M	5.1	Too big
MacBook Pro M4 Max 64GB	64 GB	Q4_K_M	3.7	Too big
RTX 4090 24GB	24 GB	Q4_K_M	3.5	Too big
RX 7900 XTX 24GB	24 GB	Q4_K_M	3.2	Too big
RTX 3090 24GB	24 GB	Q4_K_M	3.0	Too big
RTX 4080 Super 16GB	16 GB	Q4_K_M	2.8	Too big
MacBook Pro M3 Max 64GB	64 GB	Q4_K_M	2.6	Too big
MacBook Pro M1 Max 64GB	64 GB	Q4_K_M	2.4	Too big
RTX 4070 12GB	12 GB	Q4_K_M	2.0	Too big
RTX 3060 12GB	12 GB	Q4_K_M	2.0	Too big
RTX 4060 8GB	8 GB	Q4_K_M	2.0	Too big
MacBook Pro M4 Pro 48GB	48 GB	Q4_K_M	2.0	Too big

Estimates for single-stream decoding at Q4_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 Leanstral 119B A6B (119B params) fits at each quantization level on NVIDIA DGX Spark 128GB (92.2 GB usable).

Quant	Bits	VRAM	Quality	Fit
Q2_K	2	46.4 GB	Low	A84
Q3_K_S	3	58.3 GB	Low	A84
NVFP4	4	66.6 GB	Medium	A84
Q4_K_MBest for your GPU	4	72.6 GB	Medium	A84
Q5_K_M	5	85.7 GB	High	F0
Q6_K	6	97.6 GB	High	F0
Q8_0	8	127.3 GB	Very High	F0
F16	16	244.0 GB	Maximum	F0

Get started

Copy-paste commands to run Leanstral 119B A6B on your machine.

Run

docker run --rm -it ghcr.io/ggerganov/llama.cpp:full \
  --hf-repo "mistralai/Leanstral-2603" \
  --hf-file "Leanstral-2603-Q4_K_M.gguf" \
  -c 4096 -ngl 99

Your hardware

More models your NVIDIA DGX Spark 128GB can run

Model	Params	Grade	Decode
Devstral 2 123B Instruct	123B	S	2 tok/s
Qwen 3.5 122B A10B	122B	S	5 tok/s
Pixtral Large 124B	124B	A	2 tok/s

Frequently asked questions

Can NVIDIA DGX Spark 128GB run Leanstral 119B A6B?

Yes, NVIDIA DGX Spark 128GB can run Leanstral 119B A6B with a A grade (Tight fit). Expected decode speed: 5.0 tok/s.

How much VRAM does Leanstral 119B A6B need?

Leanstral 119B A6B (119B parameters) requires approximately 96.8 GB of memory with Q4_K_M quantization.

What is the best quantization for Leanstral 119B A6B?

The recommended quantization for Leanstral 119B A6B is Q4_K_M, which balances quality and memory efficiency.

What speed will Leanstral 119B A6B run at on NVIDIA DGX Spark 128GB?

On NVIDIA DGX Spark 128GB, Leanstral 119B A6B achieves approximately 5.0 tokens per second decode speed with a time-to-first-token of 38800ms using Q4_K_M quantization.

Can NVIDIA DGX Spark 128GB run Leanstral 119B A6B for coding?

For coding workloads, Leanstral 119B A6B on NVIDIA DGX Spark 128GB receives a A grade with 5.0 tok/s and 38K context.

What context window can Leanstral 119B A6B use on NVIDIA DGX Spark 128GB?

On NVIDIA DGX Spark 128GB, Leanstral 119B A6B can safely use up to 38K tokens of context. The model's official context limit is 256K, but available memory constrains the safe maximum.

What should I upgrade first if Leanstral 119B A6B feels slow on NVIDIA DGX Spark 128GB?

Prioritize bandwidth, not only capacity. If this workload feels slow, the next useful step is often a GPU tier with materially faster memory bandwidth rather than only a small bump in capacity.

Is unified memory on NVIDIA DGX Spark 128GB as fast as VRAM for Leanstral 119B A6B?

Not always. NVIDIA DGX Spark 128GB 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 NVIDIA DGX Spark 128GB See all hardware for Leanstral 119B A6B

Embed this result▼

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

<iframe src="https://willitrunai.com/embed/leanstral-119b-a6b-on-dgx-spark-128gb" width="400" height="180" frameborder="0" style="border:none;border-radius:12px;overflow:hidden;" title="Will It Run AI — fit result"></iframe>

Preview: