NVIDIA Tesla V100 SXM3 16 GB
Live eBay prices, cheapest first. The best one right now is $1,550 (open box).
Not a PCIe card. SXM module: needs an SXM-to-PCIe adapter board, its own cooling and power to run in a normal PC. SXM3 adapters are rare. The PCIe version: NVIDIA Tesla V100 PCIe 16 GB, from $325.00 →
Other sizes: 32 GB
Specs
- VRAM
- 16 GB HBM2
- Bandwidth
- 1,134 GB/s
- FP32
- 16.4 TFLOPS
- FP16
- 130 TFLOPS
- Power
- 350 W
- Connector
- SXM3 (PCIe 3.0)
- Released
- 2018
- CUDA compute
- 7.0
- Llama 3.1 8B at Q4 (est.)
- 117–175 tokens/s
Volta: no longer supported by CUDA 13; CUDA 12 builds of PyTorch, vLLM and llama.cpp still run it.
Price right now
Live on eBayFrom the live listings below; shipping included.
Listings
Live on eBay| Listing | Condition | Price | Checked |
|---|---|---|---|
| Open box | $1,550or 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- Software: CUDA 9.0 to 12.9 (not CUDA 13).
As much VRAM, about as fast, for less
- NVIDIA Tesla V100 SXM3: $599.00, 149% of the GeForce RTX 3090, 32 GB
- AMD Radeon RX 9070 XT: $890.00, 137% of the GeForce RTX 3090, 16 GB
- NVIDIA Tesla V100S PCIe: $1,099, 149% of the GeForce RTX 3090, 32 GB
Faster, for the least
- AMD Instinct MI100: $1,240, 185% of the GeForce RTX 3090, 32 GB
- NVIDIA A100 SXM4 40 GB: $2,195, 270% of the GeForce RTX 3090, 40 GB
- AMD Instinct MI210: $3,799, 211% of the GeForce RTX 3090, 64 GB
More VRAM, for not much more
- NVIDIA Tesla P40: $270.00, 35% of the GeForce RTX 3090, 24 GB
- NVIDIA GeForce RTX 3080 20 GB: $597.36, 91% of the GeForce RTX 3090, 20 GB
- NVIDIA Tesla V100 SXM3: $599.00, 149% of the GeForce RTX 3090, 32 GB
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 V100 SXM3 16 GB needed | Costs now | Speed (est.) |
|---|---|---|---|
| Llama 3.1 8B | 1 card | $1,550 | 117–175 tokens/s |
| Qwen3 14B | 1 card | $1,550 | 63–95 tokens/s |
| gpt-oss 20B | 1 card | $1,550 | 156–234 tokens/s |
| Gemma 3 27B | 2 cards | not enough listings | 34–51 tokens/s |
| Qwen3 32B | 2 cards | not enough listings | 29–43 tokens/s |
| Llama 3.3 70B | 4 cards | not enough listings | 13–20 tokens/s |
| gpt-oss 120B | 5 cards | not enough listings | 110–165 tokens/s |
| Qwen3 235B-A22B | more than 8 | — | — |
| DeepSeek R1 / V3 671B | more 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 V100 SXM3 16 GB needed | Costs now | Per question (est.) |
|---|---|---|---|
| Julia-10.144B | 1 card | $1,550 | 5.2–11 ms |
| Laya0.421B | 1 card | $1,550 | 8.6–18 ms |
| Kev-4B4B | 1 card | $1,550 | 21–44 ms |
| lev4B | 1 card | $1,550 | 62–133 ms |
| OpenJev27B | 2 cards | not enough listings | 74–159 ms |
| Clef27B | 2 cards | not enough listings | 74–159 ms |
| Perplexity Decider v1.227B | 2 cards | not enough listings | 74–159 ms |
| mammoth-127B | 2 cards | not enough listings | 74–159 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 V100 SXM3 16 GB. These open the shop's own search; we don't show their prices here.
