H200 rental price: from $3.14/hr today
The cheapest on-demand H200 today is $3.14 an hour at Prime Intellect for the H200 (SXM), checked 22 October 2026; the typical provider charges $4.59.
- Spot $2.61 Daytona, interruptible
- Marketplace $4.90 Lium, one owner's server
- Reserved $2.79 Hyperstack, term not published
- Serverless $3.00 Koyeb, per second
- In stock $5.29 RunPod, stock reported
- Provider median $4.59
- Highest on-demand $9.31
- 34 offers
- 25 providers
- Cheapest +20.8% in 90 days
- Same offers, 90 days: +6.7%
- Source primeintellect.ai
H200 offers: 34 configurations from 25 providers
More filters: provider, memory, node size, stock, place, sort, price unit
“Machines in” filters the 7 of 34 offers whose provider says where the machine is; the others are under “Place not stated”.
Cerebrium, CoreWeave, Crusoe, Daytona, DigitalOcean, fal, Fireworks AI, GMI Cloud, Hyperstack, Koyeb, Massed Compute, Modal, Nebius, OVHcloud, Spheron and Together AI put GPUs in North America on their own pages, but none of their offers here says so.
Cerebrium, CoreWeave, Crusoe, Daytona, DigitalOcean, fal, Fireworks AI, GMI Cloud, Google Cloud, Hyperstack, Koyeb, Modal, Nebius, OVHcloud, 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, Jarvislabs, Modal, RunPod and Spheron put GPUs in Asia and Oceania on their own pages, but none of their offers here says so.
- Rent →
Prime Intellect1 configuration$3.14on-demandNot reportedOn-demand $3.14
Configuration Billing $/GPU-hour Node Stock · checked H200 (SXM) on-demand $3.14 node size not published checked 22 October 2026 · source Prime Intellect H200 pricing · Prime Intellect: every GPU, fees and billing
- Rent →
GMI Cloud1 configuration$3.20on-demandNot reportedOn-demand $3.20
Configuration Billing $/GPU-hour Node Stock · checked H200 (SXM) on-demand $3.20 node size not published checked 22 October 2026 · source GMI Cloud H200 pricing · GMI Cloud: every GPU, fees and billing
- Rent →
Massed Compute1 configuration$3.62on-demandNot reportedOn-demand $3.62
Configuration Billing $/GPU-hour Node Stock · checked H200 NVL on-demand $3.62 single GPU checked 22 October 2026 · source Massed Compute H200 pricing · Massed Compute: every GPU, fees and billing
- Rent →
Hyperstack1 configuration$3.99on-demandNot reportedOn-demand $3.99Reserved $2.79
Configuration Billing $/GPU-hour Node Stock · checked H200 (SXM) reserved / committed $2.79 node size not published checked 22 October 2026 · source H200 (SXM) on-demand $3.99 node size not published checked 22 October 2026 · source Hyperstack H200 pricing · Hyperstack: every GPU, fees and billing
- Rent →
Crusoe1 configuration$4.29on-demandNot reportedOn-demand $4.29
Configuration Billing $/GPU-hour Node Stock · checked H200 (SXM) on-demand $4.29 node size not published checked 22 October 2026 · source - Rent →
DigitalOcean1 configuration$4.47on-demandNot reportedOn-demand $4.47
Configuration Billing $/GPU-hour Node Stock · checked H200 (SXM)×2 listings on-demand $4.47 single GPU checked 22 October 2026 · source DigitalOcean H200 pricing · DigitalOcean: every GPU, fees and billing
- Rent →
Daytona1 configuration$4.54on-demandNot reportedOn-demand $4.54Spot & community $2.61
Configuration Billing $/GPU-hour Node Stock · checked H200 (SXM)H200; GPU rate only, CPU/memory/disk billed separately spot / interruptible $2.61 node size not published checked 22 October 2026 · source H200 (SXM)H200; GPU rate only, CPU/memory/disk billed separately on-demand $4.54 node size not published checked 22 October 2026 · source - Rent →
Jarvislabs1 configuration$4.59on-demandNot reportedOn-demand $4.59
Configuration Billing $/GPU-hour Node Stock · checked H200 (SXM) on-demand $4.59 node size not published checked 22 October 2026 · source Jarvislabs H200 pricing · Jarvislabs: every GPU, fees and billing
- Rent →
RunPod2 configurations · H200 (SXM), H200 NVL$5.29secure cloudIn stock: lowOn-demand $5.29Spot & community $3.59Serverless $5.93
Configuration Billing $/GPU-hour Node Stock · checked H200 (SXM) community cloud $3.59 node size not published checked 22 October 2026 · source H200 (SXM) secure cloud $5.29 node size not published In stock: low · checked 22 October 2026 · source H200 (SXM)Flex worker serverless $5.93 node size not published checked 22 October 2026 · source H200 NVL secure cloud $3.79 node size not published Out of stock · checked 22 October 2026 · source - Rent →
Nebius1 configuration$5.40on-demandNot reportedOn-demand $5.40
Configuration Billing $/GPU-hour Node Stock · checked H200 (SXM) on-demand $5.40 node size not published checked 22 October 2026 · source - Rent →
Spheron1 configuration$5.51on-demandNot reportedOn-demand $5.51Spot & community $3.35
- Rent →
Verda (DataCrunch)1 configuration$5.83on-demandNot reportedOn-demand $5.83
Configuration Billing $/GPU-hour Node Stock · checked H200 (SXM) on-demand $5.83 single GPU checked 22 October 2026 · source Verda (DataCrunch) H200 pricing · Verda (DataCrunch): every GPU, fees and billing
- Rent →
Together AI1 configuration$5.99on-demandNot reportedOn-demand $5.99Spot & community $2.99
Configuration Billing $/GPU-hour Node Stock · checked H200 (SXM) spot / interruptible $2.99 node size not published checked 22 October 2026 · source H200 (SXM) on-demand $5.99 node size not published checked 22 October 2026 · source Together AI H200 pricing · Together AI: every GPU, fees and billing
- Rent →
OVHcloud1 configuration$6.20on-demandNot reportedOn-demand $6.20
Configuration Billing $/GPU-hour Node Stock · checked H200 (SXM) on-demand $6.20 8 GPUs · $49.56/hr per node checked 22 October 2026 · source OVHcloud H200 pricing · OVHcloud: every GPU, fees and billing
- Rent →
CoreWeave1 configuration$6.30on-demandNot reportedOn-demand $6.30
Configuration Billing $/GPU-hour Node Stock · checked H200 (SXM) on-demand $6.30 8 GPUs · $50.44/hr per node checked 22 October 2026 · source CoreWeave H200 pricing · CoreWeave: every GPU, fees and billing
- Rent →
AWS1 configuration$7.91on-demandNot reportedOn-demand $7.91US East (N. Virginia)
Configuration Billing $/GPU-hour Node Stock · checked H200 (SXM)p5en.48xlargeUS East (N. Virginia) on-demand $7.91 8 GPUs · $63.30/hr per node checked 22 October 2026 · source - Rent →
Google Cloud1 configuration$9.31on-demandNot reportedOn-demand $9.31Spot & community $5.59US Central (Iowa)
Configuration Billing $/GPU-hour Node Stock · checked H200 (SXM)US Central (Iowa) spot / interruptible $5.59 node size not published checked 22 October 2026 · source H200 (SXM)US Central (Iowa) on-demand $9.31 node size not published checked 22 October 2026 · source Google Cloud H200 pricing · Google Cloud: every GPU, fees and billing
- Rent →
Koyeb1 configuration$3.00serverlessNot reportedServerless $3.00
Configuration Billing $/GPU-hour Node Stock · checked H200 (SXM) serverless $3.00 node size not published checked 22 October 2026 · source - Rent →
Cerebrium1 configuration$4.20serverlessNot reportedServerless $4.20
Configuration Billing $/GPU-hour Node Stock · checked H200 (SXM) serverless $4.20 node size not published checked 22 October 2026 · source Cerebrium H200 pricing · Cerebrium: every GPU, fees and billing
- Rent →
Modal1 configuration$4.54serverlessNot reportedServerless $4.54
Configuration Billing $/GPU-hour Node Stock · checked H200 (SXM) serverless $4.54 node size not published checked 22 October 2026 · source - Rent →
Lium1 configuration$4.90marketplace hostIn stockMarketplace hosts $4.90Helsinki, Finland
Configuration Billing $/GPU-hour Node Stock · checked H200 (SXM)1-GPU hostHelsinki, Finland marketplace host $4.90 single GPU In stock · checked 22 October 2026 · source - Rent →
Replicate1 configuration$5.49serverlessNot reportedServerless $5.49
Configuration Billing $/GPU-hour Node Stock · checked H200 (SXM) serverless $5.49 node size not published checked 22 October 2026 · source Replicate H200 pricing · Replicate: every GPU, fees and billing
- Rent →
fal1 configuration$6.00serverless · checked 20 October 2026Not reportedServerless $6.00
Configuration Billing $/GPU-hour Node Stock · checked H200 (SXM) serverless $6.00 node size not published checked 20 October 2026 · stale · source - Rent →
Fireworks AI1 configuration$8.00serverlessNot reportedServerless $8.00
Configuration Billing $/GPU-hour Node Stock · checked H200 (SXM) serverless $8.00 node size not published checked 22 October 2026 · source Fireworks AI H200 pricing · Fireworks AI: every GPU, fees and billing
- Rent →
Impossible Cloud2 configurations$3.05reserved / committed · checked 17 October 2026WaitlistReserved $3.05EU; APAC
Configuration Billing $/GPU-hour Node Stock · checked H200 (SXM)313-H200; 104 GPUs quoted; 8 GPUs/node; availability by requestEU reserved / committed $3.05 node size not published Waitlist · checked 17 October 2026 · stale · source H200 (SXM)294-H200; 64 GPUs quoted; 8 GPUs/node; 8 Oct 2026APAC reserved / committed $3.16 node size not published Waitlist · checked 17 October 2026 · stale · source Impossible Cloud H200 pricing · Impossible Cloud: every GPU, fees and billing
17 providers with on-demand offers, cheapest first. Tap a provider for its configurations.
| GPU / provider | Billing | $/GPU-hour | Node · total per hour | Rent |
|---|---|---|---|---|
| H200 (SXM) · Prime Intellect141 GB | on-demand | $3.14 | – | Rent → |
| H200 (SXM) · GMI Cloud141 GB | on-demand | $3.20+23.1% in 30 days | – | Rent → |
| H200 NVL · Massed Compute141 GB | on-demand | $3.62 | 1× $3.62 | Rent → |
| H200 (SXM) · Hyperstack141 GB | on-demand | $3.99 | – | Rent →checked 22 OctDetails
|
| H200 (SXM) · Crusoe141 GB | on-demand | $4.29 | – | Rent → |
| H200 (SXM) · DigitalOcean141 GB · ×2 listings | on-demand | $4.47 | 1× $4.47 | Rent →checked 22 OctDetails
|
| H200 (SXM) · Daytona141 GB · H200; GPU rate only, CPU/memory/disk billed separately | on-demand | $4.54 | – | Rent → |
| H200 (SXM) · Jarvislabs141 GB | on-demand | $4.59+15.0% in 30 days | – | Rent → |
| H200 (SXM) · RunPod141 GB | secure cloud | $5.29+15.3% in 30 days | – | In stock: lowRent →checked 22 OctDetails
|
| H200 (SXM) · Nebius141 GB | on-demand | $5.40+20.0% in 30 days | – | Rent → |
| H200 (SXM) · Spheron141 GB | on-demand | $5.51+14.8% in 30 days | – | Rent → |
| H200 (SXM) · Verda (DataCrunch)141 GB | on-demand | $5.83+32.1% in 30 days | 1× $5.83 | Rent → |
| H200 (SXM) · Together AI141 GB | on-demand | $5.99 | – | Rent → |
| H200 (SXM) · OVHcloud141 GB | on-demand | $6.20 | 8× $49.56 | Rent → |
| H200 (SXM) · CoreWeave141 GB | on-demand | $6.30 | 8× $50.44 | Rent → |
| H200 (SXM) · AWS141 GB · p5en.48xlarge | on-demand | $7.91 | 8× $63.30 | Rent →checked 22 OctDetails
|
| H200 (SXM) · Google Cloud141 GB | on-demand | $9.31 | – | Rent →checked 22 OctDetails
|
| H200 NVL · RunPod143 GB | secure cloud | $3.79 | – | Out of stock |
| H200 (SXM) · Daytona141 GB · H200; GPU rate only, CPU/memory/disk billed separately | spot / interruptible | $2.61 | – | Rent → |
| H200 (SXM) · Hyperstack141 GB | reserved / committedTerm and minimum volume not published in this table; confirm reservation quoteStarting rate | $2.79 | – | Rent → |
| H200 (SXM) · Together AI141 GB | spot / interruptible | $2.99 | – | Rent → |
| H200 (SXM) · Koyeb141 GB | serverless | $3.00$0.000833/sec | – | Rent → |
| H200 (SXM) · Spheron141 GB | spot / interruptible | $3.35+20.9% in 30 days | – | Rent → |
| H200 (SXM) · RunPod141 GB | community cloud | $3.59 | – | Rent → |
| H200 (SXM) · Cerebrium141 GB | serverless | $4.20$0.001166/sec | – | Rent → |
| H200 (SXM) · Modal141 GB | serverless | $4.54$0.001261/sec | – | Rent →checked 22 OctDetails
|
| H200 (SXM) · Lium140 GB · 1-GPU host | marketplace host | $4.90 | 1× $4.90 | In stockRent →checked 22 OctDetails
|
| H200 (SXM) · Replicate141 GB | serverless | $5.49$0.001525/sec | – | Rent → |
| H200 (SXM) · Google Cloud141 GB | spot / interruptible | $5.59 | – | Rent →checked 22 OctDetails
|
| H200 (SXM) · RunPod141 GB · Flex worker | serverless | $5.93$0.001647/sec | – | Rent → |
| H200 (SXM) · fal141 GB | serverless | $6.00$0.001667/sec+33.3% in 30 days | – | Rent → |
| H200 (SXM) · Fireworks AI141 GB | serverless | $8.00$0.002222/sec | – | Rent → |
| H200 (SXM) · Impossible Cloud141 GB · 313-H200; 104 GPUs quoted; 8 GPUs/node; availability by request | reserved / committed12 monthsStarting rate | $3.05 | – | Waitlistchecked 17 Oct · staleDetails
|
| H200 (SXM) · Impossible Cloud141 GB · 294-H200; 64 GPUs quoted; 8 GPUs/node; 8 Oct 2026 | reserved / committed6 monthsStarting rate | $3.16 | – | Waitlistchecked 17 Oct · staleDetails
|
Try another provider or reduce the minimum memory.
Tables rank by price, lowest first. 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, "H200 rental price: from $3.14/GPU-hour", prices checked 22 October 2026, https://gpurentalprices.com/h200
Quote it with a link to this page. Downloads of the whole dataset are on Data & API.
On-demand and secure-cloud rates are capacity the provider runs, with no reservation. Spot and community capacity can be interrupted. A marketplace host is one independent owner's server listed on a marketplace. Prices are per GPU per hour, before storage and data transfer. Every row shows the date it was checked.
H200 price history and range
- Cheapest on-demand$3.14(+20.8% in 90 days)
- Provider median$4.59
- Cheapest spot$2.61
H200 price range by billing type
Per GPU-hour, today's offers. Axis $0 to $10.
CheapestProvider median (each provider's cheapest rate counted once; shown with three or more providers)
Price changes
- 30 days: +8.7% across 13 matching on-demand configurations since 22 September 2026.
- 90 days: +6.7% across 13 matching on-demand configurations since 24 July 2026.
Changes use the geometric mean of price ratios for the same provider, variant, tier, GPU count, instance and region (RunPod's place is where that day's stock sat, so it is not matched). New listings are not counted as price changes. Full ledger on the history hub.
H200 daily figures (110 days)
| Date | Cheapest on-demand | Provider median | Cheapest spot |
|---|---|---|---|
| 2026-10-22 | $3.14 | $4.59 | $2.61 |
| 2026-10-21 | $3.14 | $4.80 | $2.61 |
| 2026-10-20 | $3.14 | $4.80 | $2.61 |
| 2026-10-19 | $3.14 | $5.40 | $2.61 |
| 2026-10-18 | $3.14 | $5.40 | $2.61 |
| 2026-10-17 | $3.14 | $4.80 | $2.61 |
| 2026-10-16 | $3.14 | $5.00 | $2.61 |
| 2026-10-15 | $3.14 | $4.59 | $2.61 |
| 2026-10-14 | $3.14 | $4.59 | $2.61 |
| 2026-10-13 | $3.14 | $4.59 | $2.61 |
| 2026-10-12 | $3.14 | $4.59 | $2.61 |
| 2026-10-11 | $3.14 | $4.59 | $2.61 |
| 2026-10-10 | $3.14 | $4.59 | $2.61 |
| 2026-10-09 | $3.14 | $4.59 | $2.61 |
| 2026-10-08 | $3.14 | $4.59 | $2.61 |
| 2026-10-07 | $2.60 | $4.59 | $2.61 |
| 2026-10-06 | $2.60 | $4.59 | $2.61 |
| 2026-10-05 | $2.60 | $4.59 | $2.61 |
| 2026-10-04 | $2.60 | $4.63 | $2.99 |
| 2026-10-03 | $2.60 | $4.63 | $2.99 |
| 2026-10-02 | $2.60 | $4.47 | $2.99 |
| 2026-10-01 | $2.60 | $4.47 | $2.99 |
| 2026-09-30 | $2.60 | $4.47 | $2.99 |
| 2026-09-29 | $2.60 | $4.47 | $2.97 |
| 2026-09-28 | $2.60 | $4.47 | $2.91 |
| 2026-09-27 | $2.60 | $4.38 | $2.88 |
| 2026-09-26 | $2.60 | $4.38 | $2.90 |
| 2026-09-25 | $2.60 | $4.38 | $2.86 |
| 2026-09-24 | $2.60 | $4.38 | $2.99 |
| 2026-09-23 | $2.60 | $4.37 | $2.99 |
| 2026-09-22 | $2.60 | $4.35 | $2.77 |
| 2026-09-21 | $2.60 | $4.29 | $2.74 |
| 2026-09-20 | $2.60 | $4.33 | $2.75 |
| 2026-09-19 | $2.60 | $4.33 | $2.80 |
| 2026-09-18 | $2.60 | $4.31 | $2.77 |
| 2026-09-17 | $2.60 | $4.29 | $2.75 |
| 2026-09-16 | $2.60 | $4.29 | $2.99 |
| 2026-09-15 | $2.60 | $4.27 | $2.69 |
| 2026-09-14 | $2.60 | $4.25 | $2.69 |
| 2026-09-13 | $2.60 | $4.14 | $2.51 |
| 2026-09-12 | $2.60 | $4.14 | $2.99 |
| 2026-09-11 | $2.60 | $4.14 | $2.57 |
| 2026-09-10 | $2.60 | $4.14 | $2.99 |
| 2026-09-09 | $2.60 | $4.14 | $2.99 |
| 2026-09-08 | $2.60 | $4.14 | $2.57 |
| 2026-09-07 | $2.60 | $4.00 | $2.99 |
| 2026-09-06 | $2.60 | $4.14 | $2.99 |
| 2026-09-05 | $2.60 | $4.14 | $2.53 |
| 2026-09-04 | $2.60 | $4.14 | $2.57 |
| 2026-09-03 | $2.60 | $4.14 | $2.99 |
| 2026-09-02 | $2.60 | $4.29 | $3.31 |
| 2026-09-01 | $2.60 | $4.29 | $2.52 |
| 2026-08-31 | $2.60 | $4.29 | $2.52 |
| 2026-08-30 | $2.60 | $4.29 | $2.52 |
| 2026-08-29 | $2.60 | $4.29 | $3.31 |
| 2026-08-28 | $2.60 | $4.22 | $3.31 |
| 2026-08-27 | $2.60 | $4.22 | $3.31 |
| 2026-08-26 | $2.60 | $4.22 | $3.31 |
| 2026-08-25 | $2.60 | $4.22 | $2.53 |
| 2026-08-24 | $2.60 | $4.29 | $3.31 |
| 2026-08-23 | $2.60 | $4.22 | $2.52 |
| 2026-08-22 | $2.60 | $4.29 | $2.51 |
| 2026-08-21 | $2.60 | $4.22 | $3.31 |
| 2026-08-20 | $2.60 | $4.22 | $2.57 |
| 2026-08-19 | $2.60 | $4.22 | $3.31 |
| 2026-08-18 | $2.60 | $4.29 | $3.31 |
| 2026-08-17 | $2.60 | $4.29 | $3.31 |
| 2026-08-16 | $2.60 | $4.29 | $2.52 |
| 2026-08-15 | $2.60 | $4.29 | $2.51 |
| 2026-08-14 | $2.60 | $4.29 | $3.31 |
| 2026-08-13 | $2.60 | $4.29 | $3.31 |
| 2026-08-12 | $2.60 | $4.29 | $3.31 |
| 2026-08-11 | $2.60 | $4.29 | $2.57 |
| 2026-08-10 | $2.60 | $4.22 | $2.53 |
| 2026-08-09 | $2.60 | $4.22 | $2.52 |
| 2026-08-08 | $2.60 | $4.22 | $2.52 |
| 2026-08-07 | $2.60 | $4.22 | $2.53 |
| 2026-08-06 | $2.60 | $4.29 | $3.31 |
| 2026-08-05 | $2.60 | $4.22 | $1.76 |
| 2026-08-04 | $2.60 | $4.29 | $3.31 |
| 2026-08-03 | $2.60 | $4.29 | $1.76 |
| 2026-08-02 | $2.60 | $4.29 | $1.78 |
| 2026-08-01 | $2.60 | $4.00 | $3.31 |
| 2026-07-31 | $2.60 | $4.00 | $3.31 |
| 2026-07-30 | $2.60 | $4.00 | $3.31 |
| 2026-07-29 | $2.60 | $4.00 | $3.31 |
| 2026-07-28 | $2.60 | $4.00 | $3.31 |
| 2026-07-27 | $2.60 | $4.00 | $3.31 |
| 2026-07-26 | $2.60 | $4.00 | $3.31 |
| 2026-07-25 | $2.60 | $4.00 | $1.78 |
| 2026-07-24 | $2.60 | $4.00 | $3.31 |
| 2026-07-23 | $2.60 | $4.00 | $3.31 |
| 2026-07-22 | $2.60 | $4.00 | $3.31 |
| 2026-07-21 | $2.60 | $4.00 | $3.31 |
| 2026-07-20 | $2.60 | $4.00 | $3.31 |
| 2026-07-19 | $3.62 | $4.00 | $3.31 |
| 2026-07-18 | $3.62 | $4.29 | $3.31 |
| 2026-07-17 | $3.62 | $4.29 | $3.59 |
| 2026-07-16 | $3.62 | $4.00 | $3.59 |
| 2026-07-15 | $3.62 | $4.29 | $3.59 |
| 2026-07-14 | $3.62 | $4.29 | $3.59 |
| 2026-07-13 | $3.62 | $4.29 | $3.59 |
| 2026-07-12 | $3.62 | $4.29 | $3.59 |
| 2026-07-11 | $3.62 | $4.29 | $3.59 |
| 2026-07-10 | $3.62 | $4.29 | $3.59 |
| 2026-07-09 | $3.62 | $4.29 | $3.59 |
| 2026-07-08 | $3.62 | $4.00 | $3.59 |
| 2026-07-07 | $3.62 | $4.29 | $3.59 |
| 2026-07-06 | $3.70 | $4.00 | $3.59 |
| 2026-07-05 | $3.79 | – | $3.59 |
H200 prices by provider and billing type
Each provider's cheapest rate in each billing type, per GPU-hour. The provider name opens its H200 page with node sizes, fees and its own history.
| Provider | On-demand | Spot & community | Marketplace hosts | Reserved | Serverless |
|---|---|---|---|---|---|
| Prime Intellect | $3.14 | – | – | – | – |
| GMI Cloud | $3.20 | – | – | – | – |
| Massed Compute | $3.62 | – | – | – | – |
| RunPod | $3.79 | $3.59 | – | – | $5.93 |
| Hyperstack | $3.99 | – | – | $2.79 | – |
| Crusoe | $4.29 | – | – | – | – |
| DigitalOcean | $4.47 | – | – | – | – |
| Daytona | $4.54 | $2.61 | – | – | – |
| Jarvislabs | $4.59 | – | – | – | – |
| Nebius | $5.40 | – | – | – | – |
| Spheron | $5.51 | $3.35 | – | – | – |
| Verda (DataCrunch) | $5.83 | – | – | – | – |
| Together AI | $5.99 | $2.99 | – | – | – |
| OVHcloud | $6.20 | – | – | – | – |
| CoreWeave | $6.30 | – | – | – | – |
| AWS | $7.91 | – | – | – | – |
| Google Cloud | $9.31 | $5.59 | – | – | – |
| Koyeb | – | – | – | – | $3.00 |
| Impossible Cloudlast checked 17 October 2026 | – | – | – | $3.05 | – |
| Cerebrium | – | – | – | – | $4.20 |
| Modal | – | – | – | – | $4.54 |
| Lium | – | – | $4.90 | – | – |
| Replicate | – | – | – | – | $5.49 |
| fallast checked 20 October 2026 | – | – | – | – | $6.00 |
| Fireworks AI | – | – | – | – | $8.00 |
H200 hardware specs, by variant
| H200 (SXM) | H200 NVL | |
|---|---|---|
| Architecture | Hopper | Hopper |
| VRAM | 141 GB | 143 GB |
| FP16 TFLOPS (dense) | 989 | 835 |
| Memory bandwidth | 4,800 GB/s | 4,800 GB/s |
| TDP | 700 W | 600 W |
| Interconnect | NVLink 900GB/s | NVLink 900GB/s |
| Launched | 2023 | 2024 |
Manufacturer datasheet figures. Source: www.nvidia.com.
H200 variants
LLM memory estimates for this GPU
These models have an estimated inference footprint within 143 GB at one or more precisions. The scenario uses 4,096 tokens, batch 1, a 16-bit KV cache and 20% runtime headroom. A memory fit does not establish speed or software compatibility.
| Model | 16-bit GB | 8-bit GB | 4-bit GB |
|---|---|---|---|
| Qwen3 235B A22B | 565.2 | 283.1 | 142.0 ✓ |
| Qwen3 235B A22B 2507 | 565.2 | 283.1 | 142.0 ✓ |
| Mixtral 8x22B v0.1 | 338.6 | 169.9 | 85.5 ✓ |
| GLM 4.5 Air | 266.1 | 133.5 ✓ | 67.2 ✓ |
| Qwen2.5 72B | 176.1 | 88.9 ✓ | 45.2 ✓ |
| DeepSeek R1 Distill Llama 70B | 170.9 | 86.3 ✓ | 43.9 ✓ |
| Mixtral 8x7B v0.1 | 112.7 ✓ | 56.7 ✓ | 28.7 ✓ |
| Qwen2.5 32B | 79.9 ✓ | 40.6 ✓ | 20.9 ✓ |
| Qwen2.5 Coder 32B | 79.9 ✓ | 40.6 ✓ | 20.9 ✓ |
| DeepSeek R1 Distill Qwen 32B | 79.9 ✓ | 40.6 ✓ | 20.9 ✓ |
| QwQ 32B | 79.9 ✓ | 40.6 ✓ | 20.9 ✓ |
| Qwen3 32B | 79.9 ✓ | 40.6 ✓ | 20.9 ✓ |
| Qwen3 30B A3B | 73.8 ✓ | 37.1 ✓ | 18.8 ✓ |
| Qwen3 30B A3B 2507 | 73.8 ✓ | 37.1 ✓ | 18.8 ✓ |
| Qwen3 Coder 30B A3B | 73.8 ✓ | 37.1 ✓ | 18.8 ✓ |
| Mistral Small 3.2 24B 2506 | 58.4 ✓ | 29.6 ✓ | 15.2 ✓ |
| Mistral Small 24B 2501 | 57.4 ✓ | 29.1 ✓ | 14.9 ✓ |
| Devstral Small 2507 | 57.4 ✓ | 29.1 ✓ | 14.9 ✓ |
| Qwen2.5 14B | 36.4 ✓ | 18.7 ✓ | 9.8 ✓ |
| Qwen2.5 Coder 14B | 36.4 ✓ | 18.7 ✓ | 9.8 ✓ |
| DeepSeek R1 Distill Qwen 14B | 36.4 ✓ | 18.7 ✓ | 9.8 ✓ |
| Qwen3 14B | 36.2 ✓ | 18.5 ✓ | 9.7 ✓ |
| phi 4 | 36.2 ✓ | 18.6 ✓ | 9.8 ✓ |
| Phi 4 reasoning | 36.2 ✓ | 18.6 ✓ | 9.8 ✓ |
| Phi 4 reasoning plus | 36.2 ✓ | 18.6 ✓ | 9.8 ✓ |
| Mistral Nemo 2407 | 30.2 ✓ | 15.5 ✓ | 8.2 ✓ |
| Qwen3 8B | 20.4 ✓ | 10.6 ✓ | 5.6 ✓ |
| DeepSeek R1 Distill Llama 8B | 19.9 ✓ | 10.3 ✓ | 5.5 ✓ |
| Qwen2.5 7B | 18.6 ✓ | 9.4 ✓ | 4.9 ✓ |
| Qwen2.5 Coder 7B | 18.6 ✓ | 9.4 ✓ | 4.9 ✓ |
| DeepSeek R1 Distill Qwen 7B | 18.6 ✓ | 9.4 ✓ | 4.9 ✓ |
| Mistral 7B v0.3 | 18.0 ✓ | 9.3 ✓ | 5.0 ✓ |
| Qwen3 4B | 10.4 ✓ | 5.6 ✓ | 3.1 ✓ |
| Qwen3 4B 2507 | 10.4 ✓ | 5.6 ✓ | 3.1 ✓ |
| Qwen2.5 3B | 7.6 ✓ | 3.9 ✓ | 2.0 ✓ |
| Qwen2.5 Coder 3B | 7.6 ✓ | 3.9 ✓ | 2.0 ✓ |
| Qwen3 1.7B | 5.4 ✓ | 3.0 ✓ | 1.8 ✓ |
| DeepSeek R1 Distill Qwen 1.5B | 4.4 ✓ | 2.3 ✓ | 1.2 ✓ |
| Qwen2.5 1.5B | 3.8 ✓ | 2.0 ✓ | 1.1 ✓ |
| Qwen2.5 Coder 1.5B | 3.8 ✓ | 2.0 ✓ | 1.1 ✓ |
| Qwen3 0.6B | 2.4 ✓ | 1.5 ✓ | 1.0 ✓ |
| Qwen2.5 0.5B | 1.2 ✓ | 0.7 ✓ | 0.4 ✓ |
| Qwen2.5 Coder 0.5B | 1.2 ✓ | 0.7 ✓ | 0.4 ✓ |
Models that need more than one H200
- GLM 4.5 at 16-bit needs about 862 GB with the same context and headroom, so at least 7 cards. The cheapest H200 node with a published size that holds it: 8× at $35.76/hr (DigitalOcean, $4.47 per GPU-hour). At 8 or 4 bit it needs less; the model page recalculates.
- GLM 4.6 at 16-bit needs about 858 GB with the same context and headroom, so at least 7 cards. The cheapest H200 node with a published size that holds it: 8× at $35.76/hr (DigitalOcean, $4.47 per GPU-hour). At 8 or 4 bit it needs less; the model page recalculates.
- Qwen3 235B A22B at 16-bit needs about 565 GB with the same context and headroom, so at least 4 cards. The cheapest H200 node with a published size that holds it: 8× at $35.76/hr (DigitalOcean, $4.47 per GPU-hour). At 8 or 4 bit it needs less; the model page recalculates.
Compare model memory requirements or change context and memory assumptions on a model's page.
When to choose the H200, and what to pick instead
Start from memory: your model, its context and your batch must fit in 141 GB (the model-fit table above estimates it). Then the H200 is the right card unless one of these fits your job better.
- Pick the B200 (180 GB, from $3.49/hr) when the model or its cache does not fit in 141 GB, or finishing sooner is worth paying more: it has 2.3 times the H200's 16-bit compute.
Our workload guides recommend the H200 for LLM fine-tuning and training. Each compares it with the other cards that suit that job, at today's prices.
Answered from today's H200 prices: What is the price difference between the H100 and H200?
Rent or buy the H200
Confirmed H200 purchase listings start at $30,269 (H200 NVL 141GB, SabrePC, checked 22 October 2026, out of stock). That sum rents about 9,640 GPU-hours at today's cheapest on-demand rate of $3.14, about 4.6 years at 40 hours a week, before electricity, hosting and resale.
Compare H200 purchase prices and quote availability, then enter your usage and electricity costs in the rent-versus-buy calculator. Purchase listings have their own variants, stock status and update dates. Compare two GPUs.
H200 questions
- How much does it cost to rent an H200?
- The cheapest H200 rental we have is $3.14 per GPU-hour on Prime Intellect (on-demand), checked 22 October 2026. The median of each provider's cheapest on-demand rate is $4.59 across 17 providers.
- How much VRAM does an H200 have?
- 141-143 GB depending on the variant (H200 (SXM): 141 GB, H200 NVL: 143 GB).
- Where is the H200 cheapest right now?
- Prime Intellect has the cheapest checked on-demand rate today at $3.14 per GPU-hour; interruptible spot or community capacity starts at $2.61 (Daytona).
GPUs near the H200
- B200180 GB$3.49/hrcheapest on-demand
- MI300X192 GB$2.59/hrcheapest on-demand
- GH20096 GB$2.29/hrcheapest on-demand
- MI325X256 GB$3.80/hrcheapest on-demand
- H10080-94 GB$1.80/hrcheapest on-demand
Related
All 140 GB+ GPUs · GPUs under $2/hr · Price movers this week · Prime Intellect review · H100 vs H200 · H200 vs B200 · A100 vs H200 · Best Cloud GPU for LLM Fine-Tuning · Best Cloud GPU for LLM Training · What is the price difference between the H100 and H200? · Rubin vs H200 SXM