Loading prices…
TPU v5p rental prices
Google TPU · TPU · 95 GB HBM2e · TPU · 2023
1 providers · 2 configurations · prices as of Nov 9, 2026, 21:02 UTC
Lowest on-demand
$4.20per GPU·h · ≈ $3,066/month
at Google CloudMedian on-demand
$4.20/GPU·h
across 1 providers
Lowest spot
$0.54/GPU·h
interruptible
FP16/BF16 cost
$9.15/PFLOP·h
#66 of 69 GPUs for value
Providers
1
2 configurations
Price history
Daily TPU v5p prices per GPU-hour. The shaded period is reconstructed from public archives: the version history of SkyPilot's price catalogs, Internet Archive copies of providers' pricing pages and the open gpu-price-feed project (fewer providers than live data). After that, PetaPrice's own tracking.
Lowest on-demandMedian on-demandLowest spot
Price by provider
Cheapest On-demand TPU v5p offer at each provider, per GPU-hour. Shorter and greener is cheaper.
Google Cloud🇺🇸$4.20best
Thin vertical line = median across providers ($4.20).
Every provider we track
Pricing
Size
Host
Provider HQ
1 providers · 2 configurations
Sort
- Google CloudOn-demandSpot🇺🇸 HQ in United Statesvia SkyPilot catalogfrom · On-demandCheapest$4.20/GPU·hCheapest
What can the TPU v5p run?
Memory for the weights plus one 32K-token sequence of context, how many GPUs hold it, and what that replica costs at the cheapest in-stock on-demand rate.
| Model | Params | Memory | GPUs | $/hr | $/mo |
|---|---|---|---|---|---|
gpt-oss-20b OpenAI · 20.9B | 20.9B · 3.6B active | 20 GiB | 1 | $4.20 at Google Cloud | $3,066 |
Gemma 3 27B Google · 27.4B | 27.4B | 28 GiB | 1 | $4.20 at Google Cloud | $3,066 |
Qwen3-32B Alibaba · 32.8B | 32.8B | 39 GiB | 1 | $4.20 at Google Cloud | $3,066 |
Llama 3.3 70B Instruct Meta · 70.6B | 70.6B | 76 GiB | 1 | $4.20 at Google Cloud | $3,066 |
gpt-oss-120b OpenAI · 117B | 117B · 5.1B active | 110 GiB | 2 | $8.40 at Google Cloud | $6,132 |
Qwen3-235B-A22B Alibaba · 235B | 235B · 22B active | 225 GiB | 4 | $16.80 at Google Cloud | $12,264 |
DeepSeek-R1 DeepSeek · 684B | 684B · 37B active | 640 GiB | 8 | $33.60 at Google Cloud | $24,528 |
Kimi K3 Moonshot AI · 2.78T | 2.78T · 104B active | 2.53 TiB | 32 | $134.40 at Google Cloud | $98,112 |
Assumes 90% of each GPU's memory is usable and rounds up to 1, 2, 4, 8… GPUs for tensor parallelism (up to 64). $/hr is the GPU count times the cheapest in-stock on-demand rate per GPU; check that the provider rents instances that large. Serving many requests at once needs more memory for context.
Alternatives to the TPU v5p
GPUs with similar compute and memory, and their lowest on-demand price right now.
Gaudi 296 GB · 0.432 PF FP16$1.10/h$2.55/PFLOP·h
RTX PRO 6000 Blackwell Max-Q96 GB · 0.439 PF FP16$1.64/h$3.74/PFLOP·h- A100 SXM 80GB80 GB · 0.312 PF FP16$1.07/h$3.43/PFLOP·h
- A100 PCIe 80GB80 GB · 0.312 PF FP16$1.11/h$3.55/PFLOP·h
- RTX PRO 6000 Blackwell96 GB · 0.504 PF FP16$1.25/h$2.48/PFLOP·h
- L40S48 GB · 0.362 PF FP16$0.71/h$1.96/PFLOP·h
TPU v5p pricing questions
How much does it cost to rent an TPU v5p?
On-demand TPU v5p prices start at $4.20 per GPU-hour (Google Cloud). The median across 1 providers is $4.20, and the most expensive is $4.20. Running one GPU around the clock costs about $3,066 per month at the lowest price.
Is there cheaper spot or interruptible TPU v5p capacity?
Yes. Spot (interruptible) TPU v5p capacity starts at $0.54 per GPU-hour, 87% below the cheapest on-demand price. Spot instances can be preempted at any time, so use them for checkpointed training or batch jobs.
How much compute do you get per dollar with the TPU v5p?
At $4.20 per hour and 0.459 PF of dense FP16/BF16, the TPU v5p costs $9.15 per PFLOP-hour. That ranks 66 of 69 GPUs with on-demand prices on PetaPrice. Figures are datasheet peaks; real utilization is typically 30–60%.
What are the TPU v5p's key specs?
95 GB HBM2e with 2,765 GB/s of memory bandwidth, TPU architecture (2023), TPU form factor. Dense peak: 0.459 PF FP16 / BF16, 0.918 PF INT8.
How current are these prices?
Prices were collected on Nov 9, 2026, 21:02 UTC from provider APIs and public catalogs and refresh about every 2 hours. They are indicative list prices; always confirm on the provider's website before booking.