Specifications Compared
| Spec | GH200 | H200 |
|---|---|---|
| TDP | 900W | 700W |
| VRAM | 96 GB | 141 GB |
| CUDA Cores | 16,896 | 16,896 |
| Memory Type | HBM3 | HBM3e |
| Architecture | Hopper | Hopper |
| Form Factors | SXM | SXM, NVL |
| Interconnect | NVLink-C2C, PCIe 5.0 | 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 | 4,000 GB/s | 4,800 GB/s |
Performance Analysis
Raw compute performance matches exactly between the GH200 and H200: 1979 TFLOPS FP16 suits mixed-precision training, 3958 TFLOPS FP8 accelerates inference, and 67 TFLOPS FP32 handles general compute. This parity means similar FLOPS throughput for most deep learning tasks.
Memory capacity creates practical advantages for H200: 141 GB HBM3e versus 96 GB HBM3 enables training or inference on larger language models without excessive multi-GPU scaling. Higher 4800 GB/s bandwidth over 4000 GB/s minimizes stalls in memory-bound scenarios, supporting bigger batch sizes and faster token generation rates.
Efficiency tilts toward H200 with 700W TDP compared to 900W, allowing denser server configurations. GH200's NVLink-C2C interconnect excels in CPU-GPU unified workloads like simulations.
Live Cloud Pricing
Real-time prices from 25+ providers. Updated every 60 seconds.
GH200 Grace Hopper
| 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 | ||
 Denvr | NVIDIA GH200 Grace Hopper 96GB VRAM | 96GB | 72 vCPU 480GB RAM 7600GB Storage | Virginia | $3.87/GPU/hr | |||
 CoreWeave | NVIDIA GH200 Grace Hopper 96GB VRAM | 96GB | 72 vCPU 480GB RAM 7680GB Storage | United States | $6.50/GPU/hr |
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 GH200 Grace Hopper
The GH200 Grace Hopper stands out for workloads requiring CPU-GPU cohesion. NVLink-C2C interconnect delivers low-latency data transfer between Grace CPU and 96 GB HBM3 GPU memory, optimizing scientific simulations or hybrid applications. Its 900W TDP fits data centers equipped for high-power densities.
When to Choose the H200 SXM
The H200 proves ideal for memory-hungry AI tasks. 141 GB HBM3e VRAM accommodates massive models exceeding 96 GB, while 4800 GB/s bandwidth boosts inference throughput. At 700W TDP and pricing from $1.19/hr across 23 offers, it enables cost-effective, scalable deployments.
Use Cases
H200's 141 GB VRAM handles larger models without fragmentation, unlike GH200's 96 GB limit.
4800 GB/s bandwidth on H200 accelerates memory-bound serving with high batch sizes over GH200's 4000 GB/s.
Extra 141 GB capacity supports bigger datasets and models during fine-tuning phases.
Identical 1979 TFLOPS FP16 compute and ample VRAM above 96 GB suffice for both in image generation.
GH200's NVLink-C2C enables efficient CPU-GPU collaboration for complex simulations.
Frequently Asked Questions
What is the VRAM capacity of GH200 versus H200?▾
GH200 provides 96 GB HBM3, H200 delivers 141 GB HBM3e. Larger capacity on H200 fits bigger AI models.
Which GPU offers higher memory bandwidth?▾
H200 achieves 4800 GB/s, exceeding GH200's 4000 GB/s. This improves throughput in bandwidth-limited tasks.
How do TDP values compare?▾
GH200 requires 900W TDP, H200 uses 700W. Lower power on H200 aids rack density.
What are the current cloud prices?▾
GH200 starts at $1.99/hr average $3.59/hr over 4 offers; H200 at $1.19/hr average $3.70/hr over 23 offers.
Is compute performance identical?▾
Yes: both reach 1979 TFLOPS FP16, 67 TFLOPS FP32, 3958 TFLOPS FP8.
Which suits large LLM workloads best?▾
H200 excels with 141 GB VRAM and 4800 GB/s bandwidth for training and inference on massive models.
Which is cheaper to rent, the GH200 or the H200?▾
Cloud rental prices for both the GH200 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 GH200 have compared to the H200?▾
The GH200 has 96 GB of HBM3 memory. The H200 has 141 GB of HBM3e memory.
Can I find GH200 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 GH200 and the H200?▾
The GH200 uses the Hopper architecture (2023) while the H200 uses Hopper (2024). The H200 delivers 1.0x the FP16 throughput and 1.2x the memory bandwidth of the GH200.