PerGB

NVIDIA GeForce RTX 3090

Live eBay prices, cheapest first. The best one right now is $1,475 (used).

Specs

VRAM
24 GB GDDR6X
Bandwidth
936 GB/s
FP32
35.6 TFLOPS
FP16
71 TFLOPS
Power
350 W
Connector
PCIe 4.0
Released
2020
Launch price
$1,499
CUDA compute
8.6
Llama 3.1 8B at Q4 (est.)
96–145 tokens/s

Price right now

Live on eBay
Lowest now$1,475
Median ask$2,020
Per GB VRAM$61.46
vs launch price2% under
Live listings186

From the live listings below; shipping included.

Listings

Live on eBay
ListingConditionPriceChecked
Used$1,475−27%free shipping
Used$1,500−26%free shipping
Used$1,576or best offer−22%shipping not listed
Used$1,590or best offer−21%shipping not listed
Used$1,599−21%free shipping
Used$1,600or best offer−21%shipping not listed
Used$1,600−21%free shipping
Used$1,693or best offer−16%free shipping
Used$1,695−16%shipping not listed
Used$1,696or best offer−16%+$20.91 shipping
Used$1,700or best offer−16%shipping not listed
Used$1,700or best offer−16%shipping not listed
Used$1,700or best offer−16%shipping not listed
Used$1,700or best offer−16%shipping not listed
Used$1,710−15%free shipping
Used$1,725or best offer−15%shipping not listed
Used$1,744or best offer−14%shipping not listed
Used$1,748or best offer−13%free shipping
Used$1,749−13%shipping not listed
Used$1,750−13%shipping not listed
Used$1,757or best offer−13%free shipping
Used$1,776−12%+$28.03 shipping
Used$1,780−12%free shipping
Used$1,785or best offer−12%free shipping
Used$1,788−12%shipping not listed
Used$1,789or best offer−11%free shipping
New$1,793−11%+$4.99 shipping
Used$1,795−11%free shipping
Used$1,795−11%free shipping
Used$1,796−11%shipping not listed
New$1,799or best offer−11%free shipping
New$1,799or best offer−11%free shipping
Used$1,799or best offer−11%free shipping
New$1,799or best offer−11%free shipping
Used$1,799or best offer−11%shipping not listed
New$1,799or best offer−11%free shipping
Used$1,800or best offer−11%shipping not listed
Used$1,800−11%free shipping
Used$1,800or best offer−11%shipping not listed
Used$1,800−11%shipping not listed
Used$1,800or best offer−11%+$50.00 shipping
Used$1,800or best offer−11%shipping not listed
Used$1,800or best offer−11%shipping not listed
Used$1,825or best offer−10%shipping not listed
Used$1,850or best offer−8%free shipping
Open box$1,854or best offer−8%free shipping
Used$1,873or best offer−7%free shipping
Used$1,878or best offer−7%free shipping
Used$1,879or best offer−7%free shipping
New$1,880or best offer−7%free shipping
New$1,880or best offer−7%free shipping
Used$1,880or best offer−7%free shipping
Refurbished$1,890−6%free shipping
Open box$1,890or best offer−6%shipping not listed
Used$1,899or best offer−6%free shipping
Used$1,899−6%free shipping
Used$1,899or best offer−6%free shipping
Refurbished$1,900−6%free shipping
Refurbished$1,900−6%free shipping
Used$1,900or best offer−6%shipping not listed

Showing the 60 cheapest of 186 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 A16, vs GeForce RTX 3080 Ti, vs GeForce RTX 4070 Ti Super, vs GeForce RTX 4080.

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 3090 neededCosts nowSpeed (est.)
Llama 3.1 8B1 card$1,47596–145 tokens/s
Qwen3 14B1 card$1,47552–78 tokens/s
gpt-oss 20B1 card$1,475129–193 tokens/s
Gemma 3 27B1 card$1,47528–42 tokens/s
Qwen3 32B1 card$1,47524–35 tokens/s
Llama 3.3 70B2 cards$2,97511–16 tokens/s
gpt-oss 120B4 cards$6,14191–136 tokens/s
Qwen3 235B-A22B7 cards$10,94021–32 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 3090 neededCosts nowPer question (est.)
Julia-10.144B1 card$1,4757.7–17 ms
Laya0.421B1 card$1,47513–28 ms
Kev-4B4B1 card$1,47531–66 ms
lev4B1 card$1,47593–198 ms
OpenJev27B1 card$1,475111–237 ms
Clef27B1 card$1,475111–237 ms
Perplexity Decider v1.227B1 card$1,475111–237 ms
mammoth-127B1 card$1,475111–237 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 3090. These open the shop's own search; we don't show their prices here.