P4000
The P4000 is a cloud GPU with 8 GB of memory and 105 W TDP. It is currently available from 1 provider starting at $0.51/hr, with a market median of $0.51/hr across 1 offering.
Hardware specifications
- Architecture
- Pascal
- VRAM
- 8 GB
- Memory bandwidth
- 243 GB/s
- FP32
- 5 TFLOPS
- TDP
- 105 W
Same across all providers.
Market comparison
Offerings
No offerings found
Try clearing the spot filter or check back later.
LLMs that fit
| Model | Params | 16-bit | 8-bit | 4-bit |
|---|---|---|---|---|
Llama 3.1 70B Meta | 70B | Needs >8 GPUs | Needs >8 GPUs | Barely fits on 7× |
Qwen3.6 35B-A3B Alibaba | 35B 3B active | Needs >8 GPUs | Barely fits on 8× | 5× · 6.4 GB free |
Qwen3 32B Alibaba | 32.8B | Needs >8 GPUs | Barely fits on 7× | Barely fits on 4× |
Qwen2.5 14B Alibaba | 14B | Barely fits on 5× | 3× · 4.7 GB free | 2× · 4.6 GB free |
Llama 3.1 8B Meta | 8B | Barely fits on 3× | 2× · 4.6 GB free | 1× · 1.3 GB free |
Mistral 7B Mistral AI | 7B | 3× · 5.8 GB free | 2× · 6 GB free | 1× · 2.1 GB free |
Qwen3 4B Alibaba | 4B | 2× · 4.4 GB free | 1× · 1.6 GB free | 1× · 4.3 GB free |
Estimates: model weights plus 20% for KV cache and runtime overhead, at short context lengths. Long contexts and large batches need more. Rows marked “tight” hold the weights but not that full margin. A ×N figure is the total VRAM across N of these GPUs and makes no claim about interconnect throughput.
Similar GPUs
See more GPUs like this in
Frequently asked questions
How much does the P4000 cost per hour?
The cheapest P4000 offering starts at $0.51/hr, and the market median is $0.51/hr.
Which cloud providers offer the P4000?
No providers currently list the P4000 in our catalog.
How much VRAM does the P4000 have?
The P4000 has 8 GB of VRAM.
What LLMs can run on the P4000?
With 8 GB of VRAM, the P4000 can run 7 of our tracked models in 4-bit quantization, including Qwen3.6 35B-A3B, Qwen3 32B, Llama 3.1 70B. 4 models fit in 16-bit precision.
What's the biggest LLM I can run on multiple P4000s?
With 8× P4000 (64 GB total), the largest tracked model that fits in 16-bit is Mistral 7B (7B params), which needs 3× GPUs.