PerGB

NVIDIA Tesla P100 SXM2 12 GB

Live eBay prices, cheapest first. The best one right now is $200.00 (used).

Not a PCIe card. SXM module: needs an SXM-to-PCIe adapter board, its own cooling and power to run in a normal PC. The PCIe version: NVIDIA Tesla P100 PCIe 12 GB, from $128.00 →

Other sizes: 16 GB

Specs

VRAM
12 GB HBM2
Bandwidth
732 GB/s
FP32
10.6 TFLOPS
FP16
21.2 TFLOPS
Power
300 W
Connector
SXM2 (PCIe 3.0)
Released
2016
CUDA compute
6.0
Llama 3.1 8B at Q4 (est.)
75–113 tokens/s

Pascal: no tensor cores and no CUDA 13 support; fine for llama.cpp inference with CUDA 12 builds.

Price right now

Live on eBay
Lowest now$200.00
Median ask$200.00
Per GB VRAM$16.67
vs launch price—
Live listings1

From the live listings below; shipping included.

Listings

Live on eBay
ListingConditionPriceChecked
Used$200.00free 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

As much VRAM, about as fast, for less

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)Tesla P100 SXM2 12 GB neededCosts nowSpeed (est.)
Llama 3.1 8B1 card$200.0075–113 tokens/s
Qwen3 14B1 card$200.0041–61 tokens/s
gpt-oss 20B2 cardsnot enough listings101–151 tokens/s
Gemma 3 27B2 cardsnot enough listings22–33 tokens/s
Qwen3 32B3 cardsnot enough listings18–28 tokens/s
Llama 3.3 70B5 cardsnot enough listings8.6–13 tokens/s
gpt-oss 120B7 cardsnot enough listings71–107 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)Tesla P100 SXM2 12 GB neededCosts nowPer question (est.)
Julia-10.144B1 card$200.0016–34 ms
Laya0.421B1 card$200.0027–57 ms
Kev-4B4B1 card$200.0064–137 ms
lev4B1 card$200.00191–410 ms
OpenJev27B2 cardsnot enough listings229–490 ms
Clef27B2 cardsnot enough listings229–490 ms
Perplexity Decider v1.227B2 cardsnot enough listings229–490 ms
mammoth-127B2 cardsnot enough listings229–490 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 Tesla P100 SXM2 12 GB. These open the shop's own search; we don't show their prices here.