NVIDIA · Data centre

B200

Rented by 2 providers, though not in the same exact variant — the figures below describe different hardware and should not be read as one spread. 2 exact variants are priced separately, because they are different silicon.

Best published rate

$5.98

per GPU / hour
RunPod · 180 GB

Every published rate

cheapest first · USD per GPU per hour

Provider	What they call it	Their own price, and the arithmetic	Per GPU-hour
RunPod 180 GB	B200 NVIDIA B200	$5.98 per hour = $5.9800 per GPU-hour region Secure Cloud · 1 offer Low stock	$5.9800 $2.66 per PFLOP-hr
Vultr 192 GB	B200 vbm-256c-3072gb-8-b200-gpu	$68.00 per hour ÷ 8 GPUs = $8.5000 per GPU-hour regions not published in this feed · 1 offer	$8.5000 $3.78 per PFLOP-hr

The silicon

vendor-published

Variant	Architecture	Memory	FP16 dense	Board power
B200	Blackwell 2024	180 GB, per provider	2,250 TFLOPS	1000 W
B200	Blackwell 2024	192 GB, per provider	2,250 TFLOPS	1000 W

Vendor-published figures. FP16 throughput is dense tensor throughput with no sparsity assumed; vendors often headline the sparse figure, which is twice as large. Vendor page ↗

What they call it

Providers write the same hardware down differently. Every name below is one this index has seen for this chip; nothing was renamed to make the table tidier.

B200