Nvidia H200 SXM
The Nvidia H200 SXM is a Hopper GPU with larger and faster HBM3e memory than the H100.
- Vendor
- Nvidia
- Architecture
- Hopper
- Memory
- 141 GB HBM3e
- Memory bandwidth (TB/s)
- 4.8 TB/s
- FP8 (TFLOPS)
- 3,958 TFLOPS
- Power (W)
- 700 W
- Form factor
- SXM module (HGX H200)
- Status
- Shipping
The H200 has 141 GB of HBM3e at 4.8 TB/s and the same FP8 compute as the H100, 3,958 TFLOPS with sparsity, at up to 700 W. The SXM version ships in HGX H200 8-GPU boards; the PCIe H200 NVL card has the same 141 GB and 4.8 TB/s at up to 600 W. Nvidia states all Tensor Core figures on its H200 page with sparsity.
- Memory (GB)
- 141 GB
- Precision note (sparse or dense)
- 3,958 TFLOPS FP8 Tensor Core is stated with sparsity on Nvidia's H200 page (H200 NVL: 3,341 TFLOPS with sparsity).
- Vendor source
- nvidia.com
- Checked
Questions
How much memory does the Nvidia H200 have?
The H200 has 141 GB of HBM3e with 4.8 TB/s of memory bandwidth.
What is the difference between the H200 and H100?
The H200 is a Hopper GPU with larger and faster HBM3e memory than the H100, at the same FP8 compute.
What is the H200’s FP8 performance?
Nvidia lists 3,958 TFLOPS FP8 with sparsity for the H200 SXM and 3,341 TFLOPS for the H200 NVL.
How much power does the H200 use?
The H200 SXM is listed at up to 700 W, and the PCIe H200 NVL card at up to 600 W.
Source
Every figure is taken from the vendor’s own datasheet, product page, technical documentation or announcement, linked on each chip with the date it was last checked. How we pick and check sources: editorial standards.