GPU prices / B300-SXM-288GB

NVIDIA B300 SXM 288 GB
B300-SXM-288GB costs $4.00 to $13.80 per GPU-hour on demand across 13 cloud providers, cheapest at Lium. Prices were read from each provider's own pricing page, most recently on 2026-10-27.
Cheapest read 27 Oct · spread 3.5× · a missing provider is never an estimate
CHEAPEST $4.00/hrMEDIAN $10.28/hrSPREAD 3.5×PROVIDERS 13
Price by provider — cheapest first
Provider$/GPU·hr ↓Δ 7dTermsRead
Why a provider might be missing: its price is behind a quote form, older than 21 days, or flagged for review. A missing row is never an estimate.
Price history — cheapest daily reading
⚑ 7 of 12 weeks measuredWEEKLY LOW SHOWN
2026-08-04A FLAT MARK IS A WEEK WE DID NOT READ2026-10-20
Cost calculator
GPUs1×
Hours / month730 h
730 = 24/7 · on-demand rates only
Lium$2,920 /mo
DeepInfra$4,300 /mo
Modal$5,183 /mo
Hyperstack$5,402 /mo
Enverge$5,475 /mo
RunPod$6,563 /mo
Vast.ai$7,504 /mo
Runcrate$7,592 /mo
Verda (DataCrunch)$8,045 /mo
Lyceum$8,096 /mo
Spheron$9,023 /mo
Fal.ai$9,483 /mo
Crusoe$10,074 /mo
Spec sheet
VRAM288 GB
Bandwidth8 TB/s
FP162250 TFLOPS
FP84500 TFLOPS
TDP1400 W
NVLinkNVLink 5
ArchitectureBlackwell
Released2025
8 of 8 figures held · 6 are on record but not verified against the page they cite · only a verified figure links its source
Good for
Highest-VRAM inference part; long-context & reasoning serving at FP4 plus 70-400B training
reasoned from the specification, not measured
Benchmarks
○ Nobody has submitted this card to MLPerf, and we run no benchmarks of our own. Every submission we hold →
LLMs that fit this GPU’s VRAM
FP16
~115B params
FP8
~230B params
FP4/INT4
~460B params
Llama 3.1 405B @FP4; Qwen 72B @FP16 w/ large KV
8-GPU = 2304GB pooled; fits DeepSeek-R1 671B @FP8
ceilings at each precision, reasoned from VRAM · not a benchmark · specification unverified
Frequently asked
How much does an B300-SXM-288GB cost per hour right now?
$4.00 an hour at Lium, the cheapest of 13 sellers we read. The dearest posts $13.80 at Crusoe. Every rate links to the page it was read from, with the date it was read.
Why do B300-SXM-288GB prices vary 3.5× between providers?
Because a GPU-hour is not one product. $4.00 at Lium and $13.80 at Crusoe buy different things: interconnect, host CPU and RAM, storage, egress, support, and whether the capacity is dedicated or shared. We publish the posted rate for each and the page it came from, and do not blend them into one number.
What is the cheapest way to rent an B300-SXM-288GB long-term?
We hold no committed or reserved rate for this card — every figure here is an on-demand posted rate. Committed pricing is usually quoted rather than published, so the honest answer is that we cannot tell you, rather than a guess at a discount.
How fresh are these prices?
These rates were read between 2026-10-22 and 2026-10-27. Nothing older than 21 days is shown: a reading that goes stale is hidden rather than displayed with an old date, because a wrong price is worse than a missing one. Each row links to the page its figure came from, so any of them can be checked against the seller.
Which is cheaper for an B300-SXM-288GB — a region or a commitment?
Both move the number and they move it independently. Where a provider prices the same card in several regions we show the cheapest they charge anywhere and name the region, because a rate you can only get in Estonia is still a rate you can get. A committed term is a different contract rather than a discount on this one, so committed rates are kept out of the headline and shown separately.
Can a provider pay to be listed first for the B300-SXM-288GB?
No. The table has one stated sort at a time and nothing in it can be bought — a sponsor cannot enter it, reorder a row or change a rate. How the site is paid for, and the line each source of money does not cross, is set out under money and independence at /our-data#sec5.
Cite this page
infratoapp, “B300-SXM-288GB cloud pricing”, read 27 October 2026. $4.00 per GPU-hour at Lium. https://www.infratoapp.com/gpu/B300-SXM-288GB
Reading date travels inside the citation. How we read this data →