PerGB

NVIDIA RTX PRO 5000 Blackwell 72 GB

Live eBay prices, cheapest first. The best one right now is $8,999 (new).

Other sizes: 48 GB

Specs

VRAM
72 GB GDDR7
Bandwidth
1,344 GB/s
FP32
73 TFLOPS
FP16
293 TFLOPS
Power
300 W
Connector
PCIe 5.0
Released
2025
CUDA compute
12.0
Llama 3.1 8B at Q4 (est.)
139–208 tokens/s

Price right now

Live on eBay
Lowest now$8,999
Median ask$12,550
Per GB VRAM$124.99
vs launch price—
Live listings25

From the live listings below; shipping included.

Listings

Live on eBay
ListingConditionPriceChecked
New$8,999−28%free shipping
New$9,250−26%shipping not listed
Used$9,495or best offer−24%free shipping
New$9,500−24%free shipping
New$10,250−18%free shipping
New$10,478−17%free shipping
New$10,722or best offer−15%shipping not listed
New$10,850or best offer−14%free shipping
New$10,850or best offer−14%free shipping
New$11,181−11%free shipping
New$12,000or best offerfree shipping
Open box$12,550or best offerfree shipping
Open box$12,550or best offerfree shipping
Open box$12,550or best offerfree shipping
Open box$12,550or best offerfree shipping
Open box$12,990or best offerfree shipping
Open box$12,990or best offerfree shipping
Open box$12,990or best offerfree shipping
Open box$12,990or best offerfree shipping
Open box$12,990or best offerfree shipping
Open box$12,990or best offerfree shipping
Open box$12,990or best offerfree shipping
New$14,471or best offerfree shipping
New$17,468+$49.90 shipping
New$18,750free 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)RTX PRO 5000 Blackwell 72 GB neededCosts nowSpeed (est.)
Llama 3.1 8B1 card$8,999139–208 tokens/s
Qwen3 14B1 card$8,99975–112 tokens/s
gpt-oss 20B1 card$8,999185–277 tokens/s
Gemma 3 27B1 card$8,99940–61 tokens/s
Qwen3 32B1 card$8,99934–51 tokens/s
Llama 3.3 70B1 card$8,99916–24 tokens/s
gpt-oss 120B2 cards$18,249130–196 tokens/s
Qwen3 235B-A22B3 cards$27,74430–45 tokens/s
DeepSeek R1 / V3 671B6 cards$57,97218–27 tokens/s

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)RTX PRO 5000 Blackwell 72 GB neededCosts nowPer question (est.)
Julia-10.144B1 card$8,9993.2–6.8 ms
Laya0.421B1 card$8,9995.3–11 ms
Kev-4B4B1 card$8,99913–27 ms
lev4B1 card$8,99938–81 ms
OpenJev27B1 card$8,99945–97 ms
Clef27B1 card$8,99945–97 ms
Perplexity Decider v1.227B1 card$8,99945–97 ms
mammoth-127B1 card$8,99945–97 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 RTX PRO 5000 Blackwell 72 GB. These open the shop's own search; we don't show their prices here.