RTX PRO 5000 72GB Price & Used Prices: Which Local LLMs Can It Run?
Price alert
We'll email you when the eBay (US) price reaches your target.
The RTX PRO 5000 Blackwell 72GB, an NVIDIA workstation card shipping since late 2025, fits gpt-oss 120B alone. Confirm 72GB with nvidia-smi.
eBay
RTX PRO 5000 72GB price history
eBay
As an eBay Partner, VRAMGlass may be compensated if you buy through the eBay links on this page. The price you pay stays the same.
Full RTX PRO 5000 72GB specifications
- Memory72 GB · GDDR7 ECC
- Bandwidth1,344 GB/s
- FP16 / BF16258 TFLOPS
- FP8 / INT8516 / 516 TOPS
- Bus width
- 384-bit
- Architecture
- Blackwell (GB202)
- Cores
- 14,080
- Tensor cores
- 440
- Ecosystem
- CUDA
- Interface
- PCIe 5.0 x16
- NVLink
- Not supported
- Power
- 300 W
- Slots
- 2
- Length
- 267 mm
- Type
- Workstation
- Launch date
- Dec 18, 2025
Best RTX PRO 5000 72GB deals on eBay
Hand-picked deals from trusted eBay sellers. Shop with confidence.
As an eBay Partner, VRAMGlass may be compensated if you buy through the eBay links on this page. The price you pay stays the same.
What LLMs can a single RTX PRO 5000 72GB run?
One PRO 5000 72GB runs 70B-class and gpt-oss 120B at 4-bit, and up to 35B at 8-bit. MIG can split it into two 36GB halves for two users.
| Model | Size | Runs? | Speed t/s | Max context |
|---|---|---|---|---|
 Gemma 4 26B A4B25.2B (3.8B active) | 10.5 GB | Runs | 155.9 | 256K |
 Qwen3.6 27B27.8B | 11.8 GB | Runs | 76.3 | 256K |
| Qwen3.8 27B27.8B | 9.8 GB | Runs | 89.5 | 256K |
 GLM 4.7 Flash30B (3B active) | 11.9 GB | Runs | 164.9 | 198K |
| Gemma 4 31B30.7B | 11.8 GB | Runs | 73.7 | 256K |
| Qwen3.6 35B A3B36B (3B active) | 12.3 GB | Runs | 173.1 | 256K |
| Model | Size | Runs? | Speed t/s | Max context |
|---|---|---|---|---|
| Gemma 4 26B A4B25.2B (3.8B active) | 16.9 GB | Runs | 136 | 256K |
| Qwen3.6 27B27.8B | 16.8 GB | Runs | 55.7 | 256K |
| Qwen3.8 27B27.8B | 16.5 GB | Runs | 56.6 | 256K |
| GLM 4.7 Flash30B (3B active) | 18.3 GB | Runs | 149.5 | 198K |
| Gemma 4 31B30.7B | 18.3 GB | Runs | 50.3 | 256K |
| Qwen3.6 35B A3B36B (3B active) | 22.1 GB | Runs | 152.2 | 256K |
| Model | Size | Runs? | Speed t/s | Max context |
|---|---|---|---|---|
| Gemma 4 26B A4B25.2B (3.8B active) | 26.9 GB | Runs | 113.4 | 256K |
| Qwen3.6 27B27.8B | 28.6 GB | Runs | 34 | 256K |
| Qwen3.8 27B27.8B | 29 GB | Runs | 33.6 | 256K |
| GLM 4.7 Flash30B (3B active) | 31.8 GB | Runs | 125 | 198K |
| Gemma 4 31B30.7B | 32.6 GB | Runs | 29.6 | 256K |
| Qwen3.6 35B A3B36B (3B active) | 36.9 GB | Runs | 128.7 | 256K |
RunsFits fully in VRAM; speed is estimated from VRAM bandwidth.
Needs RAM offloadMoE only: expert weights sit in system RAM, so speed is estimated from 70 GB/s RAM bandwidth.
Too bigWon’t fit on one card at this quantization, even with RAM offload.
Full calculation rules are on the Methodology page.
What LLMs can multiple RTX PRO 5000 72GB cards run?
Two PRO 5000 72GB cards (144GB) run Qwen3.8 Flash Next at 4-bit at about 114 tokens/s; DeepSeek V4 Flash at 4-bit takes four. No NVLink.
| Model | Size | Cards | Total price | Total VRAM | Speed t/s | Total power |
|---|---|---|---|---|---|---|
| Qwen3.8 Flash Next176B (6B active) | 78.9 GB | 2 cards | $21,444 | 144 GB | 130 | 600 W |
 DeepSeek V4 Flash284B (13B active) | 96.8 GB | 2 cards | $21,444 | 144 GB | 110 | 600 W |
 Hy3295B (21B active) | 107.2 GB | 2 cards | $21,444 | 144 GB | 58.1 | 600 W |
| GLM 5.3 Flash320B (18B active) | 108.7 GB | 2 cards | $21,444 | 144 GB | 92.5 | 600 W |
| GLM 5.3744B (40B active) | 253.9 GB | 4 cards | $42,888 | 288 GB | 51.7 | 1,200 W |
| Hy4 preview770B (49B active) | 307.2 GBest. | 8 cards | $85,776 | 576 GB | 39.9 | 2,400 W |
 Muse Spark Open816B (42B active) | 311.2 GB | 8 cards | $85,776 | 576 GB | 46.6 | 2,400 W |
 Kimi K2.7 Code1,000B (32B active) | 339.5 GB | 8 cards | $85,776 | 576 GB | 61.2 | 2,400 W |
| Kimi K2.61,000B (32B active) | 340.5 GB | 8 cards | $85,776 | 576 GB | 61.1 | 2,400 W |
| Model | Size | Cards | Total price | Total VRAM | Speed t/s | Total power |
|---|---|---|---|---|---|---|
| Qwen3.8 Flash Next176B (6B active) | 111.3 GB | 2 cards | $21,444 | 144 GB | 114.1 | 600 W |
| DeepSeek V4 Flash284B (13B active) | 155.1 GB | 4 cards | $42,888 | 288 GB | 85.6 | 1,200 W |
| Hy3295B (21B active) | 182.2 GB | 4 cards | $42,888 | 288 GB | 44.7 | 1,200 W |
| GLM 5.3 Flash320B (18B active) | 199.7 GB | 4 cards | $42,888 | 288 GB | 63.4 | 1,200 W |
| GLM 5.3744B (40B active) | 467.3 GB | 8 cards | $85,776 | 576 GB | 32.8 | 2,400 W |
| Hy4 preview770B (49B active) | 467.3 GB | 8 cards | $85,776 | 576 GB | 28.6 | 2,400 W |
| Muse Spark Open816B (42B active) | 505.5 GB | 8 cards | $85,776 | 576 GB | 32.1 | 2,400 W |
| DeepSeek V4.1 Flash763B (8 / 16B active) | 510.3 GB | 8 cardsEngram in RAM | $85,776 | 576 GB | 55.5 | 2,400 W |
| Model | Size | Cards | Total price | Total VRAM | Speed t/s | Total power |
|---|---|---|---|---|---|---|
| DeepSeek V4 Flash284B (13B active) | 161.9 GB | 4 cards | $42,888 | 288 GB | 83.5 | 1,200 W |
| Qwen3.8 Flash Next176B (6B active) | 188.2 GB | 4 cards | $42,888 | 288 GB | 88.4 | 1,200 W |
| Hy3295B (21B active) | 317.7 GB | 8 cards | $85,776 | 576 GB | 31.5 | 2,400 W |
| GLM 5.3 Flash320B (18B active) | 341 GB | 8 cards | $85,776 | 576 GB | 42.6 | 2,400 W |
Rows marked “Engram in RAM” need about 203 GB of extra system RAM. Full calculation rules are on the Methodology page.
Who it suits
Good fit
- Running 70B-class dense models on one card
Llama 3.3 70B at 4-bit gets only a very short context on a 48GB card, while 72GB leaves much more room
- Running gpt-oss 120B on one card
Its expert weights ship officially in a 4-bit format (MXFP4), and 72GB holds it with room to spare
- Your system can give the GPU only 300W, but you need more than 70GB of VRAM
It is a dual-slot blower with the same power draw as the 48GB version
Not a good fit
- Running Qwen3.8 Flash Next at 2-bit on one card
It needs 78.9 GB, well over 72GB, so you need two cards or the 96GB PRO 6000
- Running only models up to 35B
Speed is exactly the same as the 48GB version, and the extra VRAM goes unused
- You want more speed
It runs the same model about 25% slower than the RTX PRO 6000
Before you buy
Make sure it's the 72GB version
Partners only began shipping the 72GB version in December 2025, and its model name and appearance are nearly identical to the 48GB version's. Don't fully trust the capacity in a listing; when the card arrives, check in nvidia-smi that total VRAM is about 72GB.
The bus is not 512-bit
Some retail pages list the 72GB version as 512-bit, which is wrong: a 512-bit bus with 3GB chips can only add up to 48GB or 96GB, never 72GB. It is actually 384-bit, with the same bandwidth as the 48GB version, so don't expect it to be faster.
MIG splits differently
MIG on the 72GB version splits it into two 36GB instances, while the 48GB version splits only into two 24GB instances. 36GB just clears the bar for 27B-class models at 8-bit: each instance can hold Qwen3.8 27B at 8-bit (29 GB), which the 48GB version's 24GB instances can't.
Compared with similar GPUs
RTX PRO 5000 72GB vs RTX PRO 5000 48GB, RTX PRO 6000, and A100 80GB PCIe: price, VRAM, bandwidth, compute and local LLM speed compared.
| Spec | RTX PRO 5000 72GB | RTX PRO 5000 48GB | RTX PRO 6000 | A100 80GB PCIe |
|---|---|---|---|---|
| Memory | 72 GB | 48 GB33% less | 96 GB33% more | 80 GB11% more |
| Bandwidth | 1,344 GB/s | 1,344 GB/sSame | 1,792 GB/s33% higher | 1,935 GB/s44% higher |
| BF16 / FP16 compute | 258 TFLOPS | 258 TFLOPSSame | 500 TFLOPS94% higher | 312 TFLOPS21% higher |
| INT8 compute | 516 TOPS | 516 TOPSSame | 1,000 TOPS94% higher | 624 TOPS21% higher |
| Qwen3.8 27B speed (2-bit) | 89.5 t/s | 89.5 t/sSame | 114.6 t/s28% faster | 122.2 t/s37% faster |
| Largest dense model (4-bit) | 122B | 79B35% less | 165B35% more | 136B11% more |
| Power | 300 W | 300 WSame | 600 W100% higher | 300 WSame |
| GPU interconnect | Not supported | Not supported | Not supported | NVLink 600 GB/s |
| Interface | PCIe 5.0 x16 | PCIe 5.0 x16 | PCIe 5.0 x16 | PCIe 4.0 x16 |
| Released | 2025-12 | 2025-03 | 2025-03 | 2021-06 |
| eBay median |
As an eBay Partner, VRAMGlass may be compensated if you buy through the eBay links on this page. The price you pay stays the same.
RTX PRO 5000 72GB rental price per hour
On Vast.ai, the RTX PRO 5000 72GB rents for $1.67 per hour on demand, about $406/month at 8 hours a day. Buying one (eBay median $10,722) pays off in about 4.5 years, 2.3 years or 9 months at 4, 8 or 24 hours a day.
| Option | Price | 4 h/day | 8 h/day | 24 h/day | View |
|---|---|---|---|---|---|
 Vast.aiRentUpdated 50 min. ago | $1.67/hr | $203/mo | $406/mo | $1,218/mo | Rent |
 eBayUpdated 8 days ago | $10,722 | Payback ~4.5 yrsPower $7/mo | Payback ~2.3 yrsPower $13/mo | Payback ~9 moPower $39/mo | Buy |
Break-even uses the US residential average ($0.18/kWh) and 300 W rated power, excluding resale value and the rest of the build.
Rental links are referral links: we may receive a commission or credit if you sign up through them. As an eBay Partner, VRAMGlass may be compensated if you buy through the eBay links on this page. The price you pay stays the same.
Amazon.jp
Xianyu