P40 vs T4: Specs, VRAM & Price
Short answer: choose the P40 when 24 GB of VRAM is the gating requirement. Choose the T4 when 16 GB is enough for the workload and availability points that way.
The P40 has 24 GB of VRAM; the T4 has 16 GB.
You can rent both by the hour, no purchase required. P40 from $0.173/GPU/hr; T4 from $0.185/GPU/hr. The P40 is 6% cheaper at the entry rate.
P40
Good for inference workloads on older-generation hardware
T4
Good for inference workloads on older-generation hardware
Last updated: 2026-10-10 22:41:56 UTCRefreshes hourly
Specs side by side
| Spec | P40 | T4 |
|---|---|---|
| VRAM | 24 GB | 16 GB |
| Memory bandwidth | – | 320+ GB/s |
| Architecture | Pascal | Turing |
| Compute capability | 6.1 | 7.5 |
| Precisions in hardware | FP32, FP16 | FP32, FP16, INT4 |
| Peak FP16/BF16 tensor TFLOPS | – | 65 TFLOPS |
| Max board power (TDP) | – | 70W |
| Form factor | PCIe | PCIe |
| Regions | 1 | 1 |
On-demand price
| Rate | P40 | T4 |
|---|---|---|
| Lowest per GPU/hr | $0.173/GPU/hr (cheapest overall) | $0.185/GPU/hr (cheapest overall) |
Which is better for AI: P40 or T4?
Choose the P40 when 24 GB of VRAM is the gating requirement. Choose the T4 when 16 GB is enough for the workload and availability points that way.
P40 vs T4: common questions
Is the P40 or the T4 better for AI?
Short answer: choose the P40 when 24 GB of VRAM is the gating requirement. Choose the T4 when 16 GB is enough for the workload and availability points that way.
Is the P40 or the T4 cheaper to rent?
On Aquanode's live marketplace the P40 starts at $0.173/GPU/hr (median $0.173/GPU/hr) and the T4 starts at $0.185/GPU/hr (median $0.185/GPU/hr). The P40 is the cheaper of the two at the entry rate, by 6%.
What is the difference between the P40 and the T4?
The P40 has 24 GB of VRAM against the T4's 16 GB; the P40 is a Pascal part and the T4 is Turing (compute capability 6.1 vs 7.5); only the T4 supports INT4 compute. On price, the P40 lists from $0.173/GPU/hr and the T4 from $0.185/GPU/hr.
How widely available are the P40 and the T4?
The P40 is listed in 1 region and the T4 in 1 region right now.
How much does 1,000 GPU-hours cost on the P40 vs the T4?
At the lowest rates listed today, 1,000 GPU-hours costs $173 on the P40 and $185 on the T4, a difference of $12 for the same runtime. Rates are per GPU per hour and update hourly.
How much VRAM does the P40 have compared to the T4?
The P40 has 24 GB of VRAM; the T4 has 16 GB.
How this comparison is calculated
Every price is normalized to a per-GPU hourly rate using the same pipeline as every other pricing surface on Aquanode, and only the cheapest qualifying offer per provider is shown. A dash means that provider does not currently list that GPU.
Architecture, compute capability and supported precisions come from each vendor's published datasheet for that generation, not from the marketplace feed. This page regenerates at most once per hour.
Memory bandwidth, tensor TFLOPS, NVLink/interconnect and max board power are the same verified-from-datasheet figures shown on each model's own P40 page and T4 page, which link the primary vendor source for that model. A dash here means Aquanode hasn't verified that figure from a primary source yet, never a guess.
Go further
Check what fits in each card's memory with the P40 VRAM calculator and the T4 VRAM calculator, see vendor-published throughput in the GPU benchmarks and specs table, or browse every model by generation in the GPU index.
More P40 comparisons
More T4 comparisons
Related reading: H100 pricing and specs, H100 vs H200, A100 vs H100, H100 and H200: SXM vs NVL vs PCIe, and The best GPUs for AI, ranked.
