GPU MODELS / P100
P100 cloud GPU pricing
4 listings across 2 providers. Cheapest published rate: $0.045/GPU/hr at QuickPod. Sorted by per-GPU hourly price. Every row links to the provider's own published price.
WHAT IS THE P100
The P100 was the first deep-learning datacenter GPU - 2016's Pascal flagship, the first with FP16 and HBM2. A decade on it survives at the bottom of catalogs: 16 GB and real bandwidth, but no tensor cores and no support for anything modern frameworks optimize for.
TENSOR COMPUTE
18.7-21.2 TFLOPS FP16 (first-generation FP16 support; no tensor cores)
INTERCONNECT
NVLink 1.0 (SXM2) or PCIe 3.0
WHAT IT IS GOOD AT
Legacy CUDA and HPC
Older codes tuned for Pascal still run, cheaply.
Teaching and coursework
16 GB for experiments at floor prices.
Modern LLM work
Not a fit - no tensor cores, no BF16, no FP8.
WHO SHOULD NOT RENT THE P100
Anything using tensor-core paths - the P100 predates them entirely.
THE P100 MARKET, BY THE NUMBERS
2 providers publish 4 P100 configurations on the panel today. Published per-GPU hourly rates run from $0.045 (QuickPod) to $0.20 (Vast.ai), a 5x spread between the cheapest and the most expensive published rate for the same chip. The median listing sits at $0.10/GPU/hr. 4 rows were reported in stock by the provider at last observation; the rest are published catalog rates. All rates are on-demand - committed-use and reserved pricing is excluded.
P100 PRICE TREND, DAILY PANEL SNAPSHOTS
Across every P100 row on the panel, the daily floor moved from $0.096 to $0.045/GPU/hr over 11 days of tracking (-52.9%). Snapshots run daily; the full history is public in the repo.
P100 PROVIDER NOTES
QuickPod · 1 config · 1 in stockSets the panel floor for this model.
from $0.045/hrVast.ai · 3 configs · 3 regions · 3 in stock3 of 3 configs reported in stock at last check.
from $0.095/hrBUYING P100: WHAT THE PANEL SAYS
At the current floor, one P100 running around the clock costs about $33 a month (730 hours of on-demand arithmetic). Per-GPU floors by config size: 1x $0.04 · 2x $0.10. Compare the alternatives below before committing - the same budget often buys more than one chip class.
P100 VS THE ALTERNATIVES
The V100 adds tensor cores and doubles everything that matters for usually a small premium. The P100 is only for code that specifically needs Pascal.
Same 16 GB, two generations newer, with tensor cores. The T4 wins for any ML use.
Compare live rates: P100 vs V100 · P100 vs T4 · P100 vs Titan Xp
P100 PRICING FAQ
What is the cheapest P100 cloud GPU right now?
The cheapest published P100 rate on the panel is $0.045 per GPU-hour at QuickPod (P100 x 1, Not published). Every price links to the provider's own published rate.
How much does a P100 cost per hour?
Across 2 providers tracked today, published per-GPU hourly rates for P100 run from $0.04 to $0.20, with a median of $0.10. Committed-use and reserved rates are excluded; everything shown is on-demand.
How many providers rent P100 GPUs?
2 providers publish 4 distinct P100 configurations on Compute Cafe. 4 of those rows were reported in stock by the provider's live endpoint at last observation.
Where is P100 cheapest?
By published per-GPU hourly rate, the cheapest geography for P100 today is united states at $0.10/hr (Vast.ai). The spread across 3 tracked geographies runs to $0.20/hr in united kingdom.
Are P100 prices going up or down?
Over the 11 days Compute Cafe has tracked the panel, the P100 floor is falling: $0.096 to $0.045/GPU/hr (-52.9%). Daily snapshots are public in the repo, so the trend is auditable.
What does a P100 cost per month?
At the current floor of $0.045/GPU/hr, a P100 running 24/7 costs about $33 per month (730 hours of on-demand arithmetic on the cheapest published rate). Committed-use contracts can price lower; the panel tracks on-demand rates only.
Is the P100 still good for anything?
Legacy HPC codes and coursework. It has 16 GB and decent bandwidth but no tensor cores - modern frameworks cannot use their fast paths on it.
See also alternatives to the P100 · cheapest GPU-hour per model · inference chips · training chips · by location
QuickPod
Vast.ai