PerGB

NVIDIA GeForce RTX 4090

Live eBay prices, cheapest first. The best one right now is $2,094 (new).

Other sizes: 48 GB

Specs

VRAM
24 GB GDDR6X
Bandwidth
1,008 GB/s
FP32
82.6 TFLOPS
FP16
165 TFLOPS
Power
450 W
Connector
PCIe 4.0
Released
2022
Launch price
$1,599
CUDA compute
8.9
Llama 3.1 8B at Q4 (est.)
104–156 tokens/s

Price right now

Live on eBay
Lowest now$2,094
Median ask$3,800
Per GB VRAM$87.25
vs launch price31% over
Live listings87

From the live listings below; shipping included.

Listings

Live on eBay
ListingConditionPriceChecked
New$2,094−45%free shipping
Used$3,024or best offer−20%+$24.02 shipping
Used$3,099−18%shipping not listed
Used$3,099−18%shipping not listed
Used$3,150−17%free shipping
Used$3,199or best offer−16%shipping not listed
Used$3,200−16%free shipping
Used$3,200−16%shipping not listed
Used$3,250or best offer−14%free shipping
Refurbished$3,299−13%free shipping
Used$3,300or best offer−13%shipping not listed
Open box$3,300or best offer−13%free shipping
Used$3,320−13%+$19.99 shipping
Used$3,334or best offer−12%free shipping
Used$3,350or best offer−12%shipping not listed
Used$3,359or best offer−12%+$59.99 shipping
Open box$3,368or best offer−11%+$68.18 shipping
Used$3,371or best offer−11%+$20.91 shipping
Used$3,400−11%free shipping
Used$3,400−11%free shipping
Used$3,400or best offer−11%shipping not listed
Open box$3,400or best offer−11%free shipping
Open box$3,400or best offer−11%shipping not listed
Used$3,400−11%free shipping
Open box$3,430or best offer−10%free shipping
Used$3,480−8%free shipping
Used$3,500−8%free shipping
Used$3,500or best offer−8%free shipping
Used$3,500or best offer−8%shipping not listed
Used$3,511or best offer−8%+$20.91 shipping
Used$3,528−7%free shipping
Used$3,555−6%+$55.00 shipping
Used$3,600or best offer−5%shipping not listed
Used$3,600−5%free shipping
New$3,605−5%+$5.00 shipping
Used$3,699free shipping
Used$3,699or best offerfree shipping
New$3,699or best offerfree shipping
Used$3,700shipping not listed
Used$3,700or best offershipping not listed
Used$3,711or best offershipping not listed
Used$3,800free shipping
Used$3,800free shipping
Used$3,800or best offerfree shipping
Used$3,920or best offershipping not listed
Used$3,924+$25.00 shipping
New$3,998or best offerfree shipping
New$3,998or best offerfree shipping
Open box$3,999shipping not listed
Refurbished$4,000free shipping
Open box$4,000or best offershipping not listed
Used$4,024or best offer+$25.00 shipping
Used$4,199or best offershipping not listed
Used$4,200or best offerfree shipping
Refurbished$4,269free shipping
New$4,399shipping not listed
Used$4,460or best offer+$60.00 shipping
Used$4,500or best offerfree shipping
Used$4,500or best offershipping not listed
New$4,550or best offerfree shipping

Showing the 60 cheapest of 87 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 PRO 4000 Blackwell, vs RTX PRO 6000 Blackwell, vs GeForce RTX 4090 D 24 GB, vs GeForce RTX 5090, vs Radeon RX 9070 XT.

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)GeForce RTX 4090 neededCosts nowSpeed (est.)
Llama 3.1 8B1 card$2,094104–156 tokens/s
Qwen3 14B1 card$2,09456–84 tokens/s
gpt-oss 20B1 card$2,094139–208 tokens/s
Gemma 3 27B1 card$2,09430–46 tokens/s
Qwen3 32B1 card$2,09425–38 tokens/s
Llama 3.3 70B2 cards$5,11812–18 tokens/s
gpt-oss 120B4 cards$11,31698–147 tokens/s
Qwen3 235B-A22B7 cards$20,86523–34 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)GeForce RTX 4090 neededCosts nowPer question (est.)
Julia-10.144B1 card$2,0944.9–10 ms
Laya0.421B1 card$2,0948.1–17 ms
Kev-4B4B1 card$2,09419–42 ms
lev4B1 card$2,09458–125 ms
OpenJev27B1 card$2,09470–150 ms
Clef27B1 card$2,09470–150 ms
Perplexity Decider v1.227B1 card$2,09470–150 ms
mammoth-127B1 card$2,09470–150 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 GeForce RTX 4090. These open the shop's own search; we don't show their prices here.