PerGB

NVIDIA L4

Live eBay prices, cheapest first. The best one right now is $3,500 (used).

Specs

VRAM
24 GB GDDR6
Bandwidth
300 GB/s
FP32
30.3 TFLOPS
FP16
121 TFLOPS
Power
72 W
Connector
PCIe 4.0
Released
2023
Launch price
$2,500 (est.)
CUDA compute
8.9
Llama 3.1 8B at Q4 (est.)
31–46 tokens/s

Running it at home

Cooling
Passive: it has no fan. In a desktop it needs air pushed through it, front to back: a fan shroud on its end (3D-printed ones with a 40 mm blower or a 92 mm fan are sold for most of these cards) or a server chassis. Without one it overheats within minutes under load.
Power
No power cable: it runs from the slot (75 W or less).
Size
Low profile, single slot.
Motherboard
Turn on "Above 4G decoding" in the BIOS: without it many boards won't start with a card with this much memory.
Screen
No display outputs: use the processor's graphics or a small second card for a monitor.

A server card: made for a rack server's airflow and power. With these sorted it works in a normal PC.

Price right now

Live on eBay
Lowest now$3,500
Median ask$4,845
Per GB VRAM$145.83
vs launch price40% over
Live listings68

From the live listings below; shipping included.

Listings

Live on eBay
ListingConditionPriceChecked
Used$3,500or best offer−28%free shipping
Used$3,700−24%free shipping
Used$3,705or best offer−24%+$66.00 shipping
Used$3,799or best offer−22%free shipping
Used$3,800or best offer−22%+$50.00 shipping
Used$3,995−18%free shipping
Used$3,995−18%free shipping
New$4,050or best offer−16%shipping not listed
Open box$4,140−15%+$10.00 shipping
Used$4,199or best offer−13%free shipping
Open box$4,310−11%+$10.00 shipping
Used$4,399or best offer−9%free shipping
Used$4,399or best offer−9%free shipping
Used$4,399or best offer−9%free shipping
Used$4,399or best offer−9%free shipping
Used$4,399or best offer−9%free shipping
Used$4,399or best offer−9%free shipping
Used$4,399−9%free shipping
Open box$4,597or best offer−5%free shipping
New$4,597or best offer−5%free shipping
New$4,667free shipping
New$4,700free shipping
New$4,718+$63.00 shipping
New$4,733+$52.00 shipping
New$4,748or best offerfree shipping
New$4,749or best offerfree shipping
New$4,750free shipping
New$4,750or best offerfree shipping
New$4,750free shipping
New$4,750free shipping
New$4,750free shipping
New$4,750free shipping
New$4,845or best offerfree shipping
New$4,845or best offerfree shipping
New$4,845or best offerfree shipping
New$4,899or best offerfree shipping
New$4,899or best offerfree shipping
New$4,917or best offerfree shipping
New$4,995free shipping
New$4,995free shipping
New$4,995free shipping
Refurbished$4,999or best offerfree shipping
Refurbished$4,999or best offerfree shipping
New$4,999or best offerfree shipping
Open box$4,999or best offerfree shipping
New$4,999or best offerfree shipping
New$4,999or best offerfree shipping
New$4,999or best offerfree shipping
New$4,999or best offerfree shipping
New$4,999or best offerfree shipping
New$4,999or best offerfree shipping
New$4,999or best offerfree shipping
New$5,000or best offerfree shipping
New$5,010+$10.00 shipping
Used$5,017or best offer+$38.88 shipping
Used$5,054or best offer+$119.90 shipping
New$5,180or best offer+$29.90 shipping
Open box$5,279or best offerfree shipping
New$5,347or best offerfree shipping
New$5,350free shipping

Showing the 60 cheapest of 68 listings.

Auctions ending soon

No auctions ending soon right now.

Prices include shipping and were read from eBay at the time shown. They can change: check the listing before you buy.

Owner reviews

No owner reviews yet. Own one? Tell others how it runs.

From signed-in owners, checked before they're published. PerGB doesn't pay for or edit reviews.

How it compares

Live on eBay

As much VRAM, about as fast, for less

Faster, for the least

More VRAM, for not much more

Compare: vs RTX 4000 Ada / SFF, vs RTX A4000, vs GeForce RTX 4070 Super.

From the live listings of every GPU, shipping included; performance against the GeForce RTX 3090 (FP16 or FP32 compute and memory bandwidth). Updated with every build.

Local AI: what it runs

Model (Q4, 8K context)L4 neededCosts nowSpeed (est.)
Llama 3.1 8B1 card$3,50031–46 tokens/s
Qwen3 14B1 card$3,50017–25 tokens/s
gpt-oss 20B1 card$3,50041–62 tokens/s
Gemma 3 27B1 card$3,5009.0–14 tokens/s
Qwen3 32B1 card$3,5007.5–11 tokens/s
Llama 3.3 70B2 cards$7,2003.5–5.3 tokens/s
gpt-oss 120B4 cards$14,70429–44 tokens/s
Qwen3 235B-A22B7 cards$26,4946.7–10 tokens/s
DeepSeek R1 / V3 671Bmore than 8——

Cost: that many of the cheapest live listings. Speed: an estimate from the memory bandwidth for one person chatting. Try other models, sizes and contexts.

Decision models

Model (Q4, 8K context)L4 neededCosts nowPer question (est.)
Julia-10.144B1 card$3,50010–22 ms
Laya0.421B1 card$3,50017–37 ms
Kev-4B4B1 card$3,50042–89 ms
lev4B1 card$3,500125–268 ms
OpenJev27B1 card$3,500150–320 ms
Clef27B1 card$3,500150–320 ms
Perplexity Decider v1.227B1 card$3,500150–320 ms
mammoth-127B1 card$3,500150–320 ms

Decision models score the options you give them instead of writing text, so they're timed per question. The time is scaled from each model's reported speed on an RTX PRO 6000 by this card's compute and memory bandwidth, within about half either way. About the models.

Compare elsewhere

Search other shops for NVIDIA L4. These open the shop's own search; we don't show their prices here.