Compact Pascal professional card for 3D visualization.
Weekly median price per GPU per hour · Get the data
By provider shows one card per company, placed where its best offer ranked. By configuration lists every offer.
What you can rent comes first: in stock, then waitlist, then not reported, then out of stock. Priced offers rank ahead of quote-only, and on-demand ahead of other billing types.
Within a group, five factors set the order:
We match the provider, its country, GPU model, form factor, billing type, availability and instance name. Matching is partial and case-insensitive. Everyday words work too, so "interruptible" finds spot and "sold out" finds out of stock.
No providers match .
Heads up: A provider's own page may quote a different figure, for example on tax, region or a promotion. A monthly-only plan shows a derived hourly rate. Verify before provisioning. More on how we price.
One P2000 has 5 GB of VRAM. Even at 4-bit, the weights of most modern AI models exceed the card's usable VRAM, and context adds KV cache on top. Below are some open-weight LLMs, with the estimated memory and GPUs each one needs.
| Model | Memory (INT4 / FP4) | P2000s needed | Cost /hr | Cost /mo |
|---|---|---|---|---|
Qwen3.8-27B
27B
|
19 GB
|
5
|
–
|
–
|
Gemma 4 31B
30.7B
|
24 GB
|
6
|
–
|
–
|
GPT-OSS-120B
117B · 5.1B active
|
68 GB
|
16
|
–
|
–
|
GLM-5.3-Flash
320B · 18B active
|
178 GB
|
40
|
–
|
–
|
MiniMax-M3
428B · 23B active
|
241 GB
|
54
|
–
|
–
|
DeepSeek V4.1 Flash
552B · 16B active
|
306 GB
|
68
|
–
|
–
|
Kimi K3
2.8T · 104B active
|
1,545 GB
|
344
|
–
|
–
|
Estimates based on the median on-demand rate. Memory is weights plus FP16/BF16 KV cache at 32K context per request. GPU counts assume 90% of advertised VRAM is usable. No guarantee of runtime support, usable performance, or that a matching quantized build exists. Pricing methodology.
Nvidia P2000 · Per GPU
| Compute · dense | |
|---|---|
| FP32 | 3 TFLOPS |
| Precision support | FP16FP32 |
| Memory | |
|---|---|
| Capacity | 5 GB GDDR5 |
| Bandwidth | 140 GB/s |
| Bus width | 160-bit |
| ECC | No |
| Silicon | |
|---|---|
| Architecture | Pascal |
| Process | TSMC 16nm FinFET |
| Transistors | 4.4 billion |
| Shader cores | 1,024 CUDA cores |
| Compute units | 8 SMs |
| Fabric and host | |
|---|---|
| Host interface | PCIe 3.0 x16 |
| Power | |
|---|---|
| Board power | 75 W |
| Cooling | Active |
Source: official Nvidia P2000 datasheet.
As of September 27, 2026, we track 2 configs from 1 provider. Prices are per GPU per hour.
| Billing type | Configs | Cheapest |
|---|---|---|
On-demand | 2 | $0.03 (in stock, Vast.ai) |
No median for on-demand: we only show this when at least 3 providers list the GPU on that billing type.
The cheapest verified in-stock on-demand P2000 tracked by GetDeploying is $0.03 per GPU per hour from Vast.ai.
The cheapest verified in-stock estimate is $22 per month on-demand.
GetDeploying currently tracks P2000 configs from 1 provider. Vast.ai has verified in-stock on-demand configs. See the full price comparison above for every provider and config.
More than one: even at 4-bit, the weights of most modern AI models exceed one P2000's usable VRAM, and context adds KV cache on top. Larger models run across multiple GPUs: the model table above shows the estimated memory and GPU count for popular open-weight LLMs.
5GB of GDDR5 in a single-slot 75W card that needs no auxiliary power connector, which is what makes it fit compact workstations and older rack chassis. Pascal still runs current CUDA builds, so it covers legacy visualization and light accelerated batch work.
5GB of VRAM and no tensor cores keep it below small-model inference, with nothing past roughly 2B parameters at 4-bit fitting in memory, and the A2000 covers the same compact-workstation slot with Ampere Tensor Cores and up to 12GB. It remains a fit for CAD, multi-display output and hardware video encode.
Nvidia RTX A2000
12 GB GDDR6 · 2.1x the bandwidth of the P2000
Nvidia P4000
8 GB GDDR5 · 74% more bandwidth than the P2000
Last updated
