Specifications Compared
| Spec | H100 | H200 |
|---|---|---|
| TDP | 700W | 700W |
| VRAM | 80-94 GB | 141 GB |
| CUDA Cores | 16,896 | 16,896 |
| Memory Type | HBM3 | HBM3e |
| Architecture | Hopper | Hopper |
| Form Factors | SXM5, PCIe, NVL | SXM, NVL |
| Interconnect | NVLink, PCIe 5.0, InfiniBand | NVLink, PCIe 5.0, InfiniBand |
| Tensor Cores | 528 | 528 |
| FP8 Performance | 3,958 TFLOPS | 3,958 TFLOPS |
| FP16 Performance | 1,979 TFLOPS | 1,979 TFLOPS |
| FP32 Performance | 67 TFLOPS | 67 TFLOPS |
| FP64 Performance | 34 TFLOPS | 34 TFLOPS |
| INT8 Performance | 3,958 TOPS | 3,958 TOPS |
| Memory Bandwidth | 3,350 GB/s | 4,800 GB/s |
Performance Analysis
Memory capacity sets H200 SXM apart: 141 GB HBM3e versus H100 PCIe 80-94 GB HBM3 enables 50-76% more model parameters in a single GPU. This supports larger batch sizes in LLM training, reducing iterations and time to convergence.
Bandwidth advantage reaches 43%: 4800 GB/s on H200 SXM over 3350 GB/s on H100 PCIe accelerates data transfers, vital for inference where high throughput lowers latency. FP16 at 1979 TFLOPS and FP32 at 67 TFLOPS remain identical, so compute-bound tasks perform equally.
In practice, H200 SXM shines in memory-bound scenarios like serving 70B+ parameter models, fitting full KV caches without quantization. H100 PCIe suffices for models under 80 GB but bottlenecks on bandwidth-intensive pipelines.
Live Cloud Pricing
Real-time prices from 25+ providers. Updated every 60 seconds.
H100 PCIe
| Provider | GPU Model | VRAM | Host Specs | Region | Price | Status | Action | |
|---|---|---|---|---|---|---|---|---|
 Hyperstack | 4×NVIDIA H100 PCIe 80GB VRAM | 80GB | 124 vCPU 720GB RAM 3300GB Storage | Canada | $1.90/GPU/hr $7.60/hr total (4×) | Available | ||
Hyperstack | 2×NVIDIA H100 PCIe 80GB VRAM | 80GB | 60 vCPU 360GB RAM 1600GB Storage | Canada | $1.90/GPU/hr $3.80/hr total (2×) | Available | ||
Hyperstack | 8×NVIDIA H100 PCIe 80GB VRAM | 80GB | 252 vCPU 1440GB RAM 6600GB Storage | Canada | $1.90/GPU/hr $15.20/hr total (8×) | Available | ||
Hyperstack | NVIDIA H100 PCIe 80GB VRAM | 80GB | 28 vCPU 180GB RAM 850GB Storage | Canada | $1.90/GPU/hr | Available | ||
Hyperstack | 8×NVIDIA H100 PCIe 80GB VRAM | 80GB | 252 vCPU 1440GB RAM 6600GB Storage | Canada | $1.95/GPU/hr $15.60/hr total (8×) | Available |
H200 SXM
| Provider | GPU Model | VRAM | Host Specs | Region | Price | Status | Action | |
|---|---|---|---|---|---|---|---|---|
 Vultr | NVIDIA GH200 Grace Hopper 96GB VRAM | 96GB | 72 vCPU 480GB RAM 960GB Storage | Atlanta | $1.99/GPU/hr | Available | ||
 Lambda Labs | NVIDIA GH200 Grace Hopper 96GB VRAM | 96GB | 64 vCPU 432GB RAM 4096GB Storage | Virginia | $2.29/GPU/hr | Available | ||
 Nebius | NVIDIA H200 SXM 141GB VRAM | 141GB | 16 vCPU 200GB RAM | 🌍Europe | $2.45/GPU/hr | |||
 CoreWeave | 8×NVIDIA H200 SXM 141GB VRAM | 141GB | 128 vCPU 0GB RAM 61440GB Storage | United States | $2.58/GPU/hr $20.64/hr total (8×) | |||
 Ori | 4×NVIDIA H200 SXM 141GB VRAM | 141GB | 96 vCPU 960GB RAM 12000GB Storage | London | $3.50/GPU/hr $14.00/hr total (4×) | Available |
When to Choose the H100 PCIe
Choose H100 PCIe for PCIe 5.0 flexibility in diverse server setups without SXM requirements. Its average cloud price of $2.77/hr beats H200 SXM's $3.70/hr, ideal for cost-sensitive projects with models fitting 80-94 GB VRAM.
Workloads like fine-tuning 13B-30B LLMs or Stable Diffusion benefit from sufficient 3350 GB/s bandwidth and lower entry pricing from $1.25/hr.
When to Choose the H200 SXM
Select H200 SXM when 141 GB HBM3e VRAM handles massive models exceeding 94 GB, such as 70B+ LLMs in training or inference. The 4800 GB/s bandwidth supports high-throughput serving in dense NVLink clusters.
High-density HPC environments favor SXM form factor despite higher average $3.70/hr pricing from $1.19/hr.
Use Cases
H200 SXM's 141 GB VRAM supports larger models and batch sizes than H100 PCIe 80-94 GB. The 4800 GB/s bandwidth reduces data bottlenecks during gradient computations.
141 GB HBM3e fits bigger KV caches for long-context inference on H200 SXM. 4800 GB/s bandwidth enables higher throughput versus H100 PCIe 3350 GB/s.
H200 SXM accommodates larger parameter counts up to 141 GB during fine-tuning. Bandwidth edge speeds optimizer steps over H100 PCIe.
H100 PCIe 80-94 GB VRAM suffices for diffusion models under 80 GB. Lower average $2.77/hr pricing fits creative workflows better than H200 SXM.
Matching 1979 TFLOPS FP16 and 67 TFLOPS FP32 suit simulations on both. Choose H100 PCIe for PCIe compatibility or H200 SXM for memory-heavy datasets.
Frequently Asked Questions
What is the VRAM difference between H100 PCIe and H200 SXM?▾
H100 PCIe offers 80-94 GB HBM3 VRAM. H200 SXM provides 141 GB HBM3e VRAM. This enables H200 SXM to load 50-76% larger models without multi-GPU sharding.
How do memory bandwidths compare?▾
H100 PCIe delivers 3350 GB/s bandwidth. H200 SXM reaches 4800 GB/s, a 43% increase. Higher bandwidth on H200 SXM accelerates inference and large-batch training.
Are compute performances identical?▾
Yes, both achieve 1979 TFLOPS FP16, 67 TFLOPS FP32, and 3958 TFLOPS FP8. Differences lie solely in memory specs and form factors.
What are current cloud prices?▾
H100 PCIe starts from $1.25/hr, average $2.77/hr across 16 offers. H200 SXM starts from $1.19/hr, average $3.70/hr across 23 offers.
Which form factor suits clusters better?▾
H200 SXM optimizes NVLink dense clusters with NVL support. H100 PCIe fits PCIe 5.0 slots for flexible, non-specialized deployments.
Does H200 SXM consume more power?▾
No, both maintain 700W TDP. Power draw matches despite H200 SXM's enhanced memory.
Which is cheaper to rent, the H100 or the H200?▾
Cloud rental prices for both the H100 and H200 vary by provider, configuration, and availability. This page shows live pricing from 25+ providers updated every 60 seconds. Scroll to the Live Cloud Pricing section to compare current rates.
How much VRAM does the H100 have compared to the H200?▾
The H100 has 80 to 94 GB of HBM3 memory. The H200 has 141 GB of HBM3e memory.
Can I find H100 and H200 GPUs available to rent right now?▾
Yes. This page shows real-time availability across 25+ cloud GPU providers. The Live Cloud Pricing section displays only in-stock offers with current pricing.
What is the main difference between the H100 and the H200?▾
The H100 uses the Hopper architecture (2022) while the H200 uses Hopper (2024). The H200 delivers 1.0x the FP16 throughput and 1.4x the memory bandwidth of the H100.