L4 cloud GPU pricing

Q: What is the cheapest L4 cloud GPU right now?

The cheapest published L4 rate on the panel is $0.080 per GPU-hour at SpotGPUs (L4 x 1, Not published). Every price links to the provider's own published rate.

Q: How much does a L4 cost per hour?

Across 26 providers tracked today, published per-GPU hourly rates for L4 run from $0.08 to $1.75, with a median of $0.70. Committed-use and reserved rates are excluded; everything shown is on-demand.

Q: How many providers rent L4 GPUs?

26 providers publish 33 distinct L4 configurations on Compute Cafe. 22 of those rows were reported in stock by the provider's live endpoint at last observation.

Q: Where is L4 cheapest?

By published per-GPU hourly rate, the cheapest geography for L4 today is india at $0.21/hr (AIC Cloud). The spread across 4 tracked geographies runs to $0.69/hr in united states.

Q: Are L4 prices going up or down?

Over the 11 days Compute Cafe has tracked the panel, the L4 floor is falling: $0.440 to $0.080/GPU/hr (-81.8%). The median continuously-tracked listing moved from $0.91 to $0.90 (-0.9%). Daily snapshots are public in the repo, so the trend is auditable.

Q: What does a L4 cost per month?

At the current floor of $0.080/GPU/hr, a L4 running 24/7 costs about $58 per month (730 hours of on-demand arithmetic on the cheapest published rate). Committed-use contracts can price lower; the panel tracks on-demand rates only.

Q: Is a full 8-GPU L4 node cheaper than a single card?

On today's panel: the single-GPU floor is $0.080/GPU/hr and the 8-GPU node floor is $0.897/GPU/hr - more expensive per GPU on a full node (+1021.8%). Check the table for the exact configs behind each floor.

Q: What size model fits on an L4?

24 GB serves a 7B at BF16, a 13B quantized, or a 34B at aggressive 4-bit quantization with tight context. It is a small-model specialist, not an H100 substitute.

33 listings across 26 providers. Cheapest published rate: $0.080/GPU/hr at SpotGPUs. Sorted by per-GPU hourly price. Every row links to the provider's own published price.

WHAT IS THE L4

The L4 is the modern successor to the T4: a 72 W single-slot card with roughly double its predecessor's throughput, FP8 support, and Ada's AV1 video engines. It is the default low-cost serving chip on every major cloud, and the right answer for small-model inference and video pipelines at scale.

MAKER

NVIDIA

ARCHITECTURE

Ada Lovelace

LAUNCHED

2023

MEMORY

24 GB GDDR6

MEMORY BANDWIDTH

300 GB/s

TENSOR COMPUTE

121 TFLOPS FP16 tensor dense (242 sparse); 242 TOPS INT8 dense

POWER DRAW

72 W

INTERCONNECT

PCIe 4.0 x16

WHAT IT IS GOOD AT

Small-model serving

7B-13B quantized models at high request rates per watt.

Video pipelines

AV1 encode/decode plus AI pre/post-processing on one low-power card.

Edge and burst inference

72 W fits dense server designs and spot-priced fleets.

WHO SHOULD NOT RENT THE L4

Models beyond ~13B - 24 GB and 300 GB/s are the ceiling.

Training of any serious size.

Provider	Config	VRAM	Location	$/hr	$/GPU/hr	Source
SpotGPUs	L4 × 1	24 GB	Not published	$0.080	$0.080	receipt
AIC Cloud	L4 × 1	24 GB	Mumbai	$0.209	$0.209	receipt
Vast.ai	L4 × 1	22 GB	South Korea, KR	$0.322	$0.322	receipt
Seeweb	L4 × 1	24 GB	Italy	$0.433	$0.433	receipt
RunPod	L4 × 1	24 GB	Global	$0.440	$0.440	receipt
JarvisLabs	L4 × 1	24 GB	India	$0.440	$0.440	receipt
gpu.ai	L4 × 1	24 GB	Not published	$0.440	$0.440	receipt
Hostnot GPU	L4 × 1	24 GB	India South	$0.490	$0.490	receipt
E2E Networks	L4 × 1	24 GB	India	$0.511	$0.511	receipt
InHosted	L4 × 1	24 GB	India	$0.511	$0.511	receipt
GPUBrazil	L4 × 1	22 GB	Asia-Pacific	$0.649	$0.649	receipt
GPUBrazil	L4 × 1	22 GB	Europe	$0.649	$0.649	receipt
UpCloud	L4 × 1	24 GB	Not published	$0.661	$0.661	receipt
Selectel	L4 × 1	24 GB	Russia	$0.667	$0.667	receipt
GPUBrazil	L4 × 1	22 GB	North America	$0.691	$0.691	receipt
Koyeb	L4 × 1	24 GB	Not published	$0.700	$0.700	receipt
Hugging Face	L4 × 1	24 GB	US	$0.700	$0.700	receipt
Google Cloud	L4 × 1	24 GB	Iowa (us-central1)	$0.707	$0.707	receipt
Lightning AI	L4 × 1	24 GB	US	$0.790	$0.790	receipt
Modal	L4 × 1	24 GB	Not published	$0.799	$0.799	receipt
Cerebrium	L4 × 1	24 GB	Not published	$0.799	$0.799	receipt
AWS	L4 × 1	24 GB	US East (N. Virginia)	$0.805	$0.805	receipt
Baseten	L4 × 1	24 GB	Not published	$0.848	$0.848	receipt
Scaleway	L4 × 4	24 GB	EU (Paris / Warsaw)	$3.590	$0.897	receipt
Scaleway	L4 × 8	24 GB	EU (Paris / Warsaw)	$7.179	$0.897	receipt
Scaleway	L4 × 1	24 GB	EU (Paris / Warsaw)	$0.900	$0.900	receipt
Scaleway	L4 × 2	24 GB	EU (Paris / Warsaw)	$1.801	$0.900	receipt
Hugging Face	L4 × 4	24 GB	US	$3.800	$0.950	receipt
Anyscale	L4 × 1	24 GB	Global	$0.954	$0.954	receipt
OVHcloud	L4 (from) × 1	24 GB	Not published	$1.000	$1.000	receipt
AWS	L4 × 4	24 GB	US East (N. Virginia)	$4.602	$1.150	receipt
Neysa	L4 (from) × 1	24 GB	India	$1.170	$1.170	receipt
Hinode	L4 (from) × 1	24 GB	Not published	$1.750	$1.750	receipt

THE L4 MARKET, BY THE NUMBERS

PROVIDERS

26

CONFIGURATIONS

33

MEDIAN $/GPU/HR

$0.70

SPREAD

21.9x

26 providers publish 33 L4 configurations on the panel today. Published per-GPU hourly rates run from $0.080 (SpotGPUs) to $1.75 (Hinode), a 22x spread between the cheapest and the most expensive published rate for the same chip. The median listing sits at $0.70/GPU/hr. 22 rows were reported in stock by the provider at last observation; the rest are published catalog rates. All rates are on-demand - committed-use and reserved pricing is excluded.

CHEAPEST L4 BY GEOGRAPHY

indiaAIC Cloud$0.209/hr south koreaVast.ai$0.322/hr italySeeweb$0.433/hr united statesGPUBrazil$0.691/hr

L4 PRICE TREND, DAILY PANEL SNAPSHOTS

Across every L4 row on the panel, the daily floor moved from $0.440 to $0.080/GPU/hr over 11 days of tracking (-81.8%). The median continuously-tracked listing held flat at $0.90. Snapshots run daily; the full history is public in the repo.

L4 PROVIDER NOTES

SpotGPUs · 1 config · 1 in stock

Sets the panel floor for this model.

from $0.080/hr

AIC Cloud · 1 config · 1 region · 1 in stock

1 of 1 configs reported in stock at last check.

from $0.209/hr

Vast.ai · 1 config · 1 region · 1 in stock

1 of 1 configs reported in stock at last check.

from $0.322/hr

Seeweb · 1 config · 1 region · 1 in stock

1 of 1 configs reported in stock at last check.

from $0.433/hr

RunPod · 1 config · 1 region · 1 in stock

1 of 1 configs reported in stock at last check.

from $0.440/hr

JarvisLabs · 1 config · 1 region

Published catalog rates.

from $0.440/hr

gpu.ai · 1 config · 1 in stock

1 of 1 configs reported in stock at last check.

from $0.440/hr

Hostnot GPU · 1 config · 1 region · 1 in stock

1 of 1 configs reported in stock at last check.

from $0.490/hr

BUYING L4: WHAT THE PANEL SAYS

At the current floor, one L4 running around the clock costs about $58 a month (730 hours of on-demand arithmetic). Node shape matters: the single-GPU floor is $0.080/GPU/hr against $0.897/GPU/hr on an 8-GPU node (+1021.8%). Per-GPU floors by config size: 1x $0.08 · 2x $0.90 · 4x $0.90 · 8x $0.90. Compare the alternatives below before committing - the same budget often buys more than one chip class.

L4 VS THE ALTERNATIVES

L4 vs T4

The L4 roughly doubles FP16 and INT8 throughput at the same 72 W and adds AV1. Unless the T4 is dramatically cheaper on the panel, take the L4.

L4 vs L40S

Triple the compute and double the memory at 5x the power and price. L4 for scale-out small models, L40S for bigger single models.

Compare live rates: L4 vs T4 · L4 vs L40S · L4 vs A10 · L4 vs A10G · L4 vs RTX A4000

L4 PRICING FAQ

What is the cheapest L4 cloud GPU right now?

The cheapest published L4 rate on the panel is $0.080 per GPU-hour at SpotGPUs (L4 x 1, Not published). Every price links to the provider's own published rate.

How much does a L4 cost per hour?

Across 26 providers tracked today, published per-GPU hourly rates for L4 run from $0.08 to $1.75, with a median of $0.70. Committed-use and reserved rates are excluded; everything shown is on-demand.

How many providers rent L4 GPUs?

26 providers publish 33 distinct L4 configurations on Compute Cafe. 22 of those rows were reported in stock by the provider's live endpoint at last observation.

Where is L4 cheapest?

By published per-GPU hourly rate, the cheapest geography for L4 today is india at $0.21/hr (AIC Cloud). The spread across 4 tracked geographies runs to $0.69/hr in united states.

Are L4 prices going up or down?

Over the 11 days Compute Cafe has tracked the panel, the L4 floor is falling: $0.440 to $0.080/GPU/hr (-81.8%). The median continuously-tracked listing moved from $0.91 to $0.90 (-0.9%). Daily snapshots are public in the repo, so the trend is auditable.

What does a L4 cost per month?

At the current floor of $0.080/GPU/hr, a L4 running 24/7 costs about $58 per month (730 hours of on-demand arithmetic on the cheapest published rate). Committed-use contracts can price lower; the panel tracks on-demand rates only.

Is a full 8-GPU L4 node cheaper than a single card?

On today's panel: the single-GPU floor is $0.080/GPU/hr and the 8-GPU node floor is $0.897/GPU/hr - more expensive per GPU on a full node (+1021.8%). Check the table for the exact configs behind each floor.

What size model fits on an L4?

24 GB serves a 7B at BF16, a 13B quantized, or a 34B at aggressive 4-bit quantization with tight context. It is a small-model specialist, not an H100 substitute.