Can InternLM 20B run on NVIDIA L4 24GB?

NO — Won't Fit

F0Won't run

○Estimated from fit model

InternLM 20B needs ~36.0 GB but NVIDIA L4 24GB only has 24.0 GB. Try a smaller quantization or lighter model.

Runtime: llama.cppCapacity: No fitBandwidth: LowStack: StandardBottleneck: Memory capacity

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

Q5_K_M (High quality) — 36.0 GB, exceeds 24.0 GB available

36.0 GB required24.0 GB available

150% VRAM needed

12.0 GB over capacity — needs offload or smaller quantization

Fit status

Too heavy

Decode

3.6 tok/s

TTFT

54187 ms

Safe context

Memory

36.0 GB / 24.0 GB

Offload

30%

Memory breakdown

Weights14.4 GB

KV Cache18.3 GB

Runtime0.9 GB

Headroom2.4 GB

See how fast it feels

With memory offload — actual speed may be lower

See how fast it feelsInternLM 20B on NVIDIA L4 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: 3.6 tok/s decode · 54.2s TTFT (warm) · 9 tok/s prefill

What limits this setup

Usable VRAM is the main blocker for this model.

Not enough usable memory

The model needs 36.0 GB, but this setup only exposes 24.0 GB of usable VRAM.

Best improvement path

Add more VRAM headroom

The first useful upgrade is more dedicated VRAM so you can fit the model without shrinking context or dropping to a much lower quant.

Performance by workload

Workload	Grade	Fit	Decode	TTFT	Context
Chat	C	Very compromised	8.2 tok/s	12915 ms	6K
Coding	F	Too heavy	4.4 tok/s	43905 ms	6K
Agentic Coding	F	Too heavy	2.1 tok/s	135916 ms	6K
Reasoning	F	Too heavy	4.4 tok/s	51887 ms	6K
RAG	F	Too heavy	2.1 tok/s	169894 ms	6K

Quantization options

How InternLM 20B (20B params) fits at each quantization level on NVIDIA L4 24GB (24.0 GB usable).

Quant	Bits	VRAM	Quality	Fit
Q2_K	2	7.8 GB	Low	B55
Q3_K_S	3	9.8 GB	Low	B57
NVFP4	4

Upgrade options

Hardware that runs InternLM 20B well

👁 NVIDIA
RTX 5090 32GBBest value

32 GB VRAM (+8)1792 GB/s (+1492)

Makes the model fit on the accelerator instead of staying completely out of reach.44.5 tok/s decode

Makes the model fit on the accelerator instead of staying completely out of reach.

Raises estimated decode speed by about 1136%.

~$1,999 MSRP

👁 NVIDIA
RTX PRO 4500 Blackwell 32GBNVIDIA upgrade

32 GB VRAM (+8)896 GB/s (+596)

Makes the model fit on the accelerator instead of staying completely out of reach.30.6 tok/s decode

Makes the model fit on the accelerator instead of staying completely out of reach.

Raises estimated decode speed by about 750%.

~$2,499 MSRP

👁 NVIDIA
RTX A6000 48GBBudget pick

48 GB VRAM (+24)768 GB/s (+468)

Makes the model fit on the accelerator instead of staying completely out of reach.41.3 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.

~$4,650 MSRP

Frequently asked questions

See all results for NVIDIA L4 24GB See all hardware for InternLM 20B

URL: https://willitrunai.com/can-run/internlm-20b-on-l4-24gb

⇱

Can InternLM 20B run on NVIDIA L4 24GB?

NO — Won't Fit

Choose the run profile you care about

Memory breakdown

See how fast it feels

What limits this setup

Best improvement path

Performance by workload

Quantization options

Hardware that runs InternLM 20B well

Frequently asked questions