PerGB

NVIDIA A2

Live eBay prices, cheapest first. The best one right now is $449.99 (refurbished).

Specs

VRAM
16 GB GDDR6
Bandwidth
200 GB/s
FP32
4.5 TFLOPS
FP16
18 TFLOPS
Power
60 W
Connector
PCIe 4.0
Released
2021
Launch price
$1,700 (est.)
CUDA compute
8.6
Llama 3.1 8B at Q4 (est.)
21–31 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$449.99
Median ask$561.40
Per GB VRAM$28.12
vs launch price74% under
Live listings130

From the live listings below; shipping included.

Listings

Live on eBay
ListingConditionPriceChecked
Refurbished$449.99or best offer−20%free shipping
Refurbished$449.99or best offer−20%free shipping
Used$479.00or best offer−15%free shipping
Used$479.00or best offer−15%free shipping
Used$479.99−15%shipping not listed
Open box$483.00−14%free shipping
Open box$483.00−14%free shipping
Open box$483.00−14%free shipping
Open box$483.00−14%free shipping
Used$483.72−14%+$7.00 shipping
New$489.00or best offer−13%free shipping
Open box$491.98−12%+$2.98 shipping
New$494.00−12%+$5.00 shipping
New$495.00or best offer−12%+$10.00 shipping
New$499.00or best offer−11%+$100.00 shipping
New$499.00or best offer−11%+$100.00 shipping
Open box$499.85or best offer−11%free shipping
New$500.00or best offer−11%free shipping
New$500.00or best offer−11%free shipping
New$500.00or best offer−11%free shipping
New$504.88−10%+$5.00 shipping
New$504.99−10%+$5.00 shipping
Used$509.99or best offer−9%free shipping
New$520.00or best offer−7%free shipping
New$520.00or best offer−7%free shipping
Used$521.40or best offer−7%free shipping
New$524.41−7%+$10.00 shipping
New$527.00or best offer−6%free shipping
Open box$528.00or best offer−6%+$75.00 shipping
New$528.00or best offer−6%free shipping
New$528.88−6%+$3.00 shipping
New$530.00or best offer−6%+$75.00 shipping
New$533.00or best offer−5%free shipping
Open box$533.50or best offerfree shipping
Used$534.98or best offerfree shipping
New$535.00free shipping
New$535.00or best offer+$50.00 shipping
Used$535.99or best offerfree shipping
Used$539.18free shipping
Used$542.84or best offerfree shipping
Used$543.00or best offerfree shipping
Used$543.84or best offerfree shipping
Open box$544.60or best offerfree shipping
Used$545.00free shipping
Used$545.84+$29.99 shipping
New$548.88+$20.00 shipping
Open box$550.00or best offerfree shipping
New$550.80+$72.00 shipping
New$550.80+$72.00 shipping
New$550.80+$72.00 shipping
Used$551.64+$10.00 shipping
Used$551.92+$120.00 shipping
Used$551.93+$125.00 shipping
Used$552.80+$79.99 shipping
Used$554.02free shipping
Used$554.02free shipping
Used$554.02free shipping
Used$555.30+$20.00 shipping
Used$555.41+$110.00 shipping
Used$556.04+$5.00 shipping

Showing the 60 cheapest of 130 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 Tesla P40, vs GeForce RTX 2060 6 GB, vs GeForce RTX 4060.

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)A2 neededCosts nowSpeed (est.)
Llama 3.1 8B1 card$449.9921–31 tokens/s
Qwen3 14B1 card$449.9911–17 tokens/s
gpt-oss 20B1 card$449.9927–41 tokens/s
Gemma 3 27B2 cards$899.986.0–9.0 tokens/s
Qwen3 32B2 cards$899.985.0–7.5 tokens/s
Llama 3.3 70B4 cards$1,8582.3–3.5 tokens/s
gpt-oss 120B5 cards$2,33819–29 tokens/s
Qwen3 235B-A22Bmore than 8——
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)A2 neededCosts nowPer question (est.)
Julia-10.144B1 card$449.9933–71 ms
Laya0.421B1 card$449.9955–118 ms
Kev-4B4B1 card$449.99133–284 ms
lev4B1 card$449.99398–852 ms
OpenJev27B2 cards$899.98475–1018 ms
Clef27B2 cards$899.98475–1018 ms
Perplexity Decider v1.227B2 cards$899.98475–1018 ms
mammoth-127B2 cards$899.98475–1018 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 A2. These open the shop's own search; we don't show their prices here.