PerGB

NVIDIA GeForce RTX 3080 20 GB

Live eBay prices, cheapest first. The best one right now is $597.36 (open box).

Other sizes: 12 GB, 10 GB

Specs

VRAM
20 GB GDDR6X
Bandwidth
912 GB/s
FP32
29.8 TFLOPS
FP16
60 TFLOPS
Power
320 W
Connector
PCIe 4.0
Released
2020
CUDA compute
8.6
Llama 3.1 8B at Q4 (est.)
94–141 tokens/s

Price right now

Live on eBay
Lowest now$597.36
Median ask$975.00
Per GB VRAM$29.87
vs launch price—
Live listings21

From the live listings below; shipping included.

Listings

Live on eBay
ListingConditionPriceChecked
Open box$597.36−39%free shipping
Used$908.98or best offer−7%+$120.00 shipping
Open box$916.00or best offer−6%+$98.00 shipping
Open box$941.36or best offer+$5.51 shipping
Open box$949.55or best offerfree shipping
Used$950.00or best offershipping not listed
Open box$966.90or best offerfree shipping
Open box$969.00or best offerfree shipping
Open box$975.00or best offerfree shipping
Open box$975.00or best offerfree shipping
Open box$975.00or best offerfree shipping
Open box$979.00or best offerfree shipping
Open box$980.00free shipping
Open box$1,000or best offerfree shipping
Open box$1,005+$2.97 shipping
Open box$1,027+$49.99 shipping
Open box$1,030or best offerfree shipping
Open box$1,030or best offerfree shipping
New$1,162or best offerfree shipping
New$1,173or best offerfree shipping
New$1,191or best offerfree shipping

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

Faster, for the least

More VRAM, for not much more

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 3080 20 GB neededCosts nowSpeed (est.)
Llama 3.1 8B1 card$597.3694–141 tokens/s
Qwen3 14B1 card$597.3651–76 tokens/s
gpt-oss 20B1 card$597.36125–188 tokens/s
Gemma 3 27B2 cards$1,50627–41 tokens/s
Qwen3 32B2 cards$1,50623–34 tokens/s
Llama 3.3 70B3 cards$2,42211–16 tokens/s
gpt-oss 120B4 cards$3,36488–133 tokens/s
Qwen3 235B-A22B8 cards$7,19921–31 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 3080 20 GB neededCosts nowPer question (est.)
Julia-10.144B1 card$597.368.5–18 ms
Laya0.421B1 card$597.3614–30 ms
Kev-4B4B1 card$597.3634–73 ms
lev4B1 card$597.36102–219 ms
OpenJev27B1 card$597.36122–261 ms
Clef27B1 card$597.36122–261 ms
Perplexity Decider v1.227B1 card$597.36122–261 ms
mammoth-127B1 card$597.36122–261 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 3080 20 GB. These open the shop's own search; we don't show their prices here.