Can EXAONE 4.0 32B run on NVIDIA A10 24GB?

Q: What should I upgrade first if EXAONE 4.0 32B feels slow on NVIDIA A10 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.

BARELY — Tight on Memory

A73Great

Estimated from fit model

EXAONE 4.0 32B needs ~26.7 GB VRAM. NVIDIA A10 24GB has 24.0 GB. With Q4_K_M quantization, expect ~16 tok/s.

Runtime: llama.cppCapacity: 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.

Capabilities:

Select quantization to explore

Q4_K_M (Medium quality) — 26.7 GB, 15.5 tok/s, Very compromised (needs ~2 GB host RAM)

26.7 GB required24.0 GB available

111% VRAM needed

2.7 GB over capacity — needs offload or smaller quantization

Fit status

Very compromised (needs ~2 GB host RAM)

Decode

15.5 tok/s

TTFT

12503 ms

Safe context

Memory

26.7 GB / 24.0 GB

Offload

10%

Memory breakdown

Weights19.5 GB

KV Cache3.9 GB

Runtime0.9 GB

Headroom2.4 GB

See how fast it feels

See how fast it feelsEXAONE 4.0 32B on NVIDIA A10 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: 15.5 tok/s decode · 12.5s TTFT (warm) · 39 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 10% 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.

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 2.0 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	A	Runs with offload (needs ~0.6 GB host RAM)	18.2 tok/s	5813 ms	5K
Coding	A	Very compromised (needs ~2 GB host RAM)	15.5 tok/s	12503 ms	5K
Agentic Coding	F	Too heavy	11.6 tok/s	24235 ms	5K
Reasoning	A	Very compromised (needs ~2 GB host RAM)	15.5 tok/s	14776 ms	5K
RAG	F	Too heavy	11.6 tok/s	30294 ms

Quantization options

How EXAONE 4.0 32B (32B params) fits at each quantization level on NVIDIA A10 24GB (24.0 GB usable).

Quant	Bits	VRAM	Quality	Fit
Q2_K	2	12.5 GB	Low	S85
Q3_K_S	3	15.7 GB	Low	A85
NVFP4Best for your GPU

Get started

Copy-paste commands to run EXAONE 4.0 32B on your machine.

Run

ollama run exaone-4:32b

Your hardware

More models your NVIDIA A10 24GB can run

Model	Params	Grade	Decode	Capabilities
Qwen 3.6 35B A3B	35B	A	30.5 tok/s

Frequently asked questions

Can NVIDIA A10 24GB run EXAONE 4.0 32B?

Yes, NVIDIA A10 24GB can run EXAONE 4.0 32B with a A grade (Very compromised (needs ~2 GB host RAM)). Expected decode speed: 15.5 tok/s.

How much VRAM does EXAONE 4.0 32B need?

EXAONE 4.0 32B (32B parameters) requires approximately 26.7 GB of memory with Q4_K_M quantization.

What is the best quantization for EXAONE 4.0 32B?

The recommended quantization for EXAONE 4.0 32B is Q4_K_M, which balances quality and memory efficiency.

What speed will EXAONE 4.0 32B run at on NVIDIA A10 24GB?

On NVIDIA A10 24GB, EXAONE 4.0 32B achieves approximately 15.5 tokens per second decode speed with a time-to-first-token of 12503ms using Q4_K_M quantization.

Can NVIDIA A10 24GB run EXAONE 4.0 32B for coding?

For coding workloads, EXAONE 4.0 32B on NVIDIA A10 24GB receives a A grade with 15.5 tok/s and 5K context.

What context window can EXAONE 4.0 32B use on NVIDIA A10 24GB?

On NVIDIA A10 24GB, EXAONE 4.0 32B can safely use up to 5K tokens of context. The model's official context limit is 131K, but available memory constrains the safe maximum.

What should I upgrade first if EXAONE 4.0 32B feels slow on NVIDIA A10 24GB?

See all results for NVIDIA A10 24GB See all hardware for EXAONE 4.0 32B

Embed this result▼

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

<iframe src="https://willitrunai.com/embed/exaone-4-32b-on-a10-24gb" width="400" height="180" frameborder="0" style="border:none;border-radius:12px;overflow:hidden;" title="Will It Run AI — fit result"></iframe>

Preview: