Rubin vs B200
Rubin has 288 GB of HBM4 at 22 TB/s per GPU, against 180 GB of HBM3e at 8 TB/s on B200. On demand, B200 rents from $5.00/GPU-hour and Rubin from $24.36/GPU-hour (21 November 2026; 21 November 2026).
Specifications, side by side
| Specification | Rubin | B200 |
|---|---|---|
| Exact GPU | Rubin | B200 |
| Memory per GPU | 288 GB | 180 GB |
| Memory generation | HBM4 | HBM3e |
| Memory bandwidth per GPU | 22 TB/s | 8 TB/s ¹ |
| NVLink bandwidth per GPU | 3.6 TB/s | 1.8 TB/s |
| FP16/BF16 compute per GPU, dense | 4 PFLOPS | 2.25 PFLOPS ² |
| Availability | Announced; NVIDIA says partners offer it from the second half of 2026 · 1 recorded offers | Rentable now: 37 offers from 21 providers (21 November 2026) |
Rubin has 1.60 times the published memory capacity. Calculation: 288 ÷ 180. Specifications checked 5 October 2026.
These are NVIDIA's published peak figures, not measured speed on any workload.
¹ Our calculation. Calculation: 64 TB/s ÷ 8 GPUs (NVIDIA DGX B200).
² Our calculation. Calculation: 36 PFLOPS sparse ÷ 2 ÷ 8 GPUs (NVIDIA HGX B200/B300 table).
Recorded on-demand rental rates
| Rate | Rubin | B200 |
|---|---|---|
| Lowest recorded on-demand | $24.36/GPU-hour (recorded 21 November 2026) | $5.00/GPU-hour (recorded 21 November 2026) |
Every rate comes from our daily price collection and shows the date it was recorded.
Rubin offers and billing conditions
More filters: provider, memory, node size, stock, sort, price unit
| GPU / provider | Billing | $/GPU-hour | Node · total per hour | Rent |
|---|---|---|---|---|
| Rubin · OVHcloud288 GB | on-demand | $24.36 | 8× $194.89 | Rent → |
Try another provider or reduce the minimum memory.
Fresh prices come first. Within each billing type, prices run lowest first, with unavailable offers last. Commercial relationships never affect ordering. How we collect and rank prices.
Same GPU, provider, billing and price on several hosts or listings: one row, with the count. – means the provider does not state it.
Monthly view assumes 730 hours of continuous use. Daily view assumes 24 hours.
Cite this data
GPU Rental Prices, "Rubin vs B200: GPU memory and rental prices", prices checked 21 November 2026, https://gpurentalprices.com/vs/rubin-vs-b200
Quote it with a link to this page. Downloads of the whole dataset are on Data & API.
Current B200 rental offers
More filters: provider, memory, node size, stock, place, CPU and RAM, sort, price unit
“Machines in” filters the 9 of 36 offers whose provider says where the machine is; the others are under “Place not stated”.
“vCPUs per GPU” and “RAM per GPU” use the 5 of 36 offers whose provider publishes the machine's vCPUs or RAM for a stated number of GPUs (Lambda and RunPod); an offer that does not publish them never matches a minimum. A GiB counts as 1.07 GB.
Baseten, Cerebrium, CoreWeave, Daytona, fal, Fireworks AI, GMI Cloud, Hyperstack, Koyeb, Lambda, Modal, Nebius, Spheron and Together AI put GPUs in North America on their own pages, but none of their offers here says so.
Baseten, Cerebrium, CoreWeave, Daytona, fal, Fireworks AI, GMI Cloud, Google Cloud, Hyperstack, Koyeb, Lambda, Modal, Nebius, Spheron, Together AI and Verda (DataCrunch) put GPUs in Europe on their own pages, but none of their offers here says so.
Fireworks AI, GMI Cloud, Google Cloud, Lambda, Modal, RunPod and Spheron put GPUs in Asia and Oceania on their own pages, but none of their offers here says so.
| GPU / provider | Billing | $/GPU-hour | Node · total per hour | Rent |
|---|---|---|---|---|
| B200 · GMI Cloud180 GB | on-demand | $5.00 | – | Rent → |
| B200 · Hyperstack192 GB | on-demand | $6.00 | – | Rent →checked 21 NovDetails
|
| B200 · Daytona180 GB · B200; GPU rate only, CPU/memory/disk billed separately | on-demand | $7.49+19.8% in 30 days | – | Rent → |
| B200 · Lambda180 GB · 8x NVIDIA HGX B200 | on-demand | $7.69 | 8× $61.52 | Rent →checked 21 NovDetails
|
| B200 · Lambda180 GB · 4x NVIDIA B200 SXM6 | on-demand | $7.79+14.7% in 30 days | 4× $31.16 | Rent →checked 21 NovDetails
|
| B200 · Lambda180 GB · 2x NVIDIA B200 SXM6 | on-demand | $7.89+14.5% in 30 days | 2× $15.78 | Rent →checked 21 NovDetails
|
| B200 · RunPod180 GB | secure cloud | $7.99 | – | In stock: lowRent →checked 21 NovDetails
|
| B200 · Lambda180 GB · 1x NVIDIA B200 SXM6 | on-demand | $7.99+14.3% in 30 days | 1× $7.99 | Rent →checked 21 NovDetails
|
| B200 · Nebius180 GB | on-demand | $8.50 | – | Rent → |
| B200 · CoreWeave180 GB | on-demand | $8.60 | 8× $68.80 | Rent → |
| B200 · Together AI180 GB | on-demand | $9.29+13.4% in 30 days | – | Rent → |
| B200 · Verda (DataCrunch)180 GB | on-demand | $9.75+17.8% in 30 days | 1× $9.75 | Rent → |
| B200 · Spheron192 GB | on-demand | $10.25−7.8% in 30 days | – | Rent → |
| B200 · Oracle Cloud180 GB · BM.GPU.B200.8; bare metal, 8 GPUs; billed per second after the first hour | on-demand | $14.00 | 8× $112.00 | Rent →checked 21 NovDetails
|
| B200 · Google Cloud180 GB | on-demand | $16.11 | – | Rent →checked 21 NovDetails
|
| B200 · Daytona180 GB · B200; GPU rate only, CPU/memory/disk billed separately | spot / interruptible | $4.30+19.8% in 30 days | – | Rent → |
| B200 · Together AI180 GB | spot / interruptible | $4.65+13.7% in 30 days | – | Rent → |
| B200 · Verda (DataCrunch)180 GB | spot / interruptible | $4.87 | 1× $4.87 | Rent → |
| B200 · Lium179 GB · ×2 hosts | spot / interruptible | $5.00 | 1× $5.00 | In stockRent →checked 21 NovDetails
|
| B200 · Hyperstack180 GB | reserved / committedTerm and minimum volume not published in this table; confirm reservation quoteStarting rate | $5.10 | – | Rent → |
| B200 · Spheron192 GB | spot / interruptible | $5.25−8.1% in 30 days | – | Rent → |
| B200 · Google Cloud180 GB | spot / interruptible | $5.45+10.0% in 30 days | – | Rent →checked 21 NovDetails
|
| B200 · Koyeb180 GB | serverless | $5.50$0.001528/sec | – | Rent → |
| B200 · RunPod180 GB | community cloud | $5.98 | – | Rent → |
| B200 · Cerebrium180 GB | serverless | $6.01$0.001670/sec | – | Rent → |
| B200 · Modal180 GB | serverless | $6.25$0.001736/sec | – | Rent →checked 21 NovDetails
|
| B200 · Lium179 GB · 1-GPU host | marketplace host | $6.45 | 1× $6.45 | In stockRent →checked 21 NovDetails
|
| B200 · Lium179 GB · 1-GPU host | marketplace host | $6.70 | 1× $6.70 | In stockRent →checked 21 NovDetails
|
| B200 · fal192 GB | serverless | $7.99$0.002219/sec | – | Rent → |
| B200 · Lium179 GB · 8-GPU host | marketplace host | $8.08 | 8× $64.64 | In stockRent →checked 21 NovDetails
|
| B200 · Lium179 GB · 8-GPU host | marketplace host | $8.89 | 8× $71.12 | In stockRent →checked 21 NovDetails
|
| B200 · Baseten180 GB | serverless | $9.98$0.002772/sec | – | Rent → |
| B200 · RunPod180 GB · Flex worker | serverless | $10.79$0.002997/sec+24.9% in 30 days | – | Rent → |
| B200 · Fireworks AI180 GB | serverless | $13.00$0.003611/sec | – | Rent → |
| B200 · Vultr192 GB | spot / interruptible | $3.20 | 8× $25.60 | Out of stock |
| B200 · Impossible Cloud180 GB · b200-768gpu-599-b200; delivery Immediate | reserved / committed36 monthsStarting rate | $6.23 | – | Waitlist |
Try another provider or reduce the minimum memory.
A spot machine can be stopped whenever the provider needs it back: how much notice each provider gives, and what happens to the disk.
Fresh prices come first. Within each billing type, prices run lowest first, with unavailable offers last. Commercial relationships never affect ordering. How we collect and rank prices.
Same GPU, provider, billing and price on several hosts or listings: one row, with the count. – means the provider does not state it.
Monthly view assumes 730 hours of continuous use. Daily view assumes 24 hours.
Cite this data
GPU Rental Prices, "Rubin vs B200: GPU memory and rental prices", prices checked 21 November 2026, https://gpurentalprices.com/vs/rubin-vs-b200
Quote it with a link to this page. Downloads of the whole dataset are on Data & API.
Specifications and sources
Rubin values come from NVIDIA's individual GPU specification table. B200 capacity comes from its NVIDIA specification. Checked 5 October 2026. Quoted evidence is listed under Sources in the footer.
B200 memory type · NVIDIA DGX B200 · NVIDIA HGX B200/B300 · Blackwell Ultra specifications · NVIDIA H200
NVIDIA publishes two sets of per-GPU Rubin figures. Its HGX table lists 22 TB/s of memory bandwidth and 3.6 TB/s of NVLink bandwidth; its Vera Rubin NVL72 table lists 19.2 TB/s and 3 TB/s, for an at-scale AI factory. Ask the provider which system it rents before comparing performance.
B200 rental prices and history · Rubin specifications and cloud availability · All GPU comparisons