Data updated 2026-11-04 15:10 UTC · refreshed every 5 min
Current Price
$0.16
+0.5% (24h)
Available Now
3 GPU
Limited
Total in Network
15 GPU
Live data
Utilization
80.0%
High
Market history
USD/hour (minimum · median)
Ready to rent
Infrastructure Growth
Estimated with our VRAM calculator for a single card: 8K context, 4 concurrent users, vLLM-style serving.
21B MoE · 3.6B active
~32 tok/s per stream
Where the memory goes
9B
~31 tok/s per stream
Where the memory goes
4B
~29 tok/s per stream
Where the memory goes
Estimates, not benchmarks: real usage depends on the runtime, context and batch settings.
Memory
24GB GDDR5
FP32 Performance
12.0 TFLOPS
Tensor Performance
N/A
Memory Bandwidth
346 GB/s
Architecture
Pascal
TDP
250W
FP64 Performance
0.35 TFLOPS
CUDA Cores
3840
The Tesla P40 was designed for inference workloads with 24GB memory and INT8 performance.
The NVIDIA Tesla P40 features 24GB of GDDR5 memory with 346 GB/s bandwidth, making it excellent for ai inference, video transcoding, deep learning. With N/A tensor performance and 12.0 TFLOPS FP32 compute, this NVIDIA Pascal GPU delivers exceptional performance for demanding workloads.
Currently, 15 GPU Tesla P40 GPUs are available in the GPUniq network with 80.0% utilization. The average rental price is $0.16 per hour, with 3 GPU ready for immediate deployment.
Prices are live minimum / median USD per hour on the GPUniq marketplace, as of 2026-11-04 15:10 UTC.
| GPU | VRAM | Architecture | FP32 | Tensor | Bandwidth | TDP | Min $/hr | Median $/hr | Available |
|---|---|---|---|---|---|---|---|---|---|
| Tesla P40 | 24GB GDDR5 | Pascal | 12.0 TFLOPS | N/A | 346 GB/s | 250W | $0.16 | $0.16 | 4 |
| Tesla V100 | 32GB HBM2 | Volta | 15.7 TFLOPS | 125 TFLOPS | 900 GB/s | 300W | $0.07 | $0.12 | 1472 |
| Tesla T4 | 16GB GDDR6 | Turing | 8.1 TFLOPS | 65 TFLOPS | 320 GB/s | 70W | $0.20 | $0.20 | 1 |
| Tesla P100 | 16GB HBM2 | Pascal | 9.3 TFLOPS | N/A | 732 GB/s | 250W | — | — | 0 |
The current rental price for NVIDIA Tesla P40 is $0.16 per hour on GPUniq. Prices may vary based on demand and availability.
NVIDIA Tesla P40 is ideal for AI Inference, Video Transcoding, Deep Learning. With 24GB of GDDR5 memory and N/A tensor performance, it excels at demanding AI and compute workloads.
Currently, there are 3 GPU available for immediate rental out of 15 GPU total in the GPUniq network. The current utilization is 80.0%.
On a single NVIDIA Tesla P40, our VRAM calculator fits GPT-OSS 20B (INT4 / AWQ, 14.2 of 24 GB); Qwen3.5-9B (INT4 / AWQ, 10.8 of 24 GB); Qwen3.5-4B (FP16, 11.4 of 24 GB) — estimated at 8K context with 4 concurrent users.
NVIDIA Tesla P40 features 24GB GDDR5 memory, 12.0 TFLOPS FP32 performance, N/A tensor performance, 346 GB/s memory bandwidth, and 250W TDP. It uses the Pascal architecture.
NVIDIA official specificationsAPI docsAbout GPUniqContact supportPrivacy policyTerms of use
$0.16/hr
Tesla P40 · 3 GPU