Price check
Compare normalized hourly prices, quote freshness, and the minimum number of GPUs required by each provider.
Use the current provider quotes on this page to compare B300 hourly cost, availability, and update time before budgeting a large AI deployment.
Compare all GPU rental prices โPrices updated:
Starting from
$4.9900
per GPU / hour
Prices are updated every hour via automated scrapers. Sorted by lowest price.
| Platform | Price/Hour (USD) | Price/Hour (CNY) | Availability | Action |
|---|---|---|---|---|
Prime IntellectCheapest | $4.9900 | ยฅ36.18 | High | Rent โ |
vast.ai | $6.7049 | ยฅ48.61 | Low | Rent โ |
RunPod | $6.9400 | ยฅ50.32 | High | Rent โ |
Modal | $7.1000 | ยฅ51.47 | High | Rent โ |
Hyperstack | $7.4000 | ยฅ53.65 | High | Rent โ |
DataCrunch | $7.5000 | ยฅ54.38 | High | Rent โ |
Scaleway | $8.6800 | ยฅ62.93 | High | Rent โ |
Daily average price per platform over the past 7 days.
B300 planning should begin with current per-GPU pricing and node requirements because provider quotes can represent either one GPU or a multi-GPU configuration.
Compare normalized hourly prices, quote freshness, and the minimum number of GPUs required by each provider.
Use B300 for workloads whose scale and memory requirements justify current-generation infrastructure; smaller jobs may be cheaper on H100 or B200.
Open the B200 and H100 pages to compare current market cost and provider coverage before committing to B300.
NVIDIA B300 rental currently starts from $4.9900/hour on Prime Intellect. Prices can change with inventory; check the page update time.
NVIDIA B300 has 288GB VRAM, making it ideal for LLM inference, AI training, and other high-performance computing workloads.
Among currently tracked quotes, Prime Intellect has the lowest NVIDIA B300 price at $4.9900/hour.
Compare per-GPU hourly price, provider availability, region, minimum rental configuration, and quote freshness. Confirm whether a listed offer is priced per GPU or per multi-GPU node.
The comparison aims to normalize prices per GPU, but provider minimums can still be based on multi-GPU nodes. Verify the provider terms before purchase.
Continue comparing