GPU comparison
H200 vs MI325X: specs and rental prices
NVIDIA H200 against AMD Instinct MI325X: 141 GB against 256 GB of memory, lowest listed on-demand price $3.99 against $3.80 per GPU-hour. Specs from NVIDIA and AMD, prices as of Oct 25, 2026.
In short
- The MI325X lists 256 GB of GPU memory against 141 GB for the H200: 1.82× higher.
- Memory type is HBM3e on the H200 and HBM3E on the MI325X.
- The MI325X lists 6,000 GB/s of memory bandwidth against 4,800 GB/s for the H200: 1.25× higher.
- The MI325X lists 2,610 TFLOPS of FP16 / BF16 tensor throughput against 1,979 TFLOPS for the H200: 1.32× higher.
- Maximum power is listed at 700 W for the H200 and 1,000 W for the MI325X.
- Not every figure can be compared: at least one vendor page does not publish: GPU-to-GPU interconnect.
- The lowest listed on-demand price is $3.80 per GPU-hour for the MI325X (DigitalOcean) and $3.99 for the H200 (Hyperstack).
- Per GB of GPU memory that is $0.028 per hour for the H200 and $0.015 for the MI325X.
This summary is generated from the tables below. Prices were last checked Oct 25, 2026. A listing not re-checked for 7 days is greyed and left out of every lowest-price figure; after 14 days it is removed from the page.
Specifications side by side
| Specification | H200 SXM | MI325X | Difference |
|---|---|---|---|
| GPU memory | 141 GB | 256 GB | MI325X 1.82× higher |
| Memory type | HBM3e | HBM3E | Different |
| Memory bandwidth | 4,800 GB/s | 6,000 GB/s | MI325X 1.25× higher |
| FP8 tensor throughput, with sparsity | 3,958 TFLOPS | 5,220 TFLOPS | MI325X 1.32× higher |
| FP16 / BF16 tensor throughput, with sparsity | 1,979 TFLOPS | 2,610 TFLOPS | MI325X 1.32× higher |
| TF32 tensor throughput, with sparsity | 989 TFLOPS | 1,300 TFLOPS | MI325X 1.31× higher |
| FP64 throughput | 34 TFLOPS | 81.7 TFLOPS | MI325X 2.4× higher |
| Maximum power | 700 W | 1,000 W | MI325X 1.43× higher |
| GPU-to-GPU interconnect | 900 GB/s | Not published | — |
Compared versions: H200 SXM and MI325X. The figure in the stronger cell is shaded; power is shown without a winner. Two throughput figures are compared only when both vendors state them on the same basis (both with sparsity, or both dense), and “derived” means a total the vendor publishes divided by the GPU count it publishes. Full tables with sources: H200 specifications · MI325X specifications.
Rental prices
| Rental price | H200 | MI325X | Difference |
|---|---|---|---|
| Lowest listed on-demand price | $3.99 per GPU-hour · Hyperstack (SXM) · checked Oct 25, 2026 | $3.80 per GPU-hour · DigitalOcean · checked Oct 25, 2026 | H200 1.05× higher |
| Lowest listed spot price | $0.79 per GPU-hour · Nebius · checked Oct 25, 2026 | No fresh listing | — |
| Price per GB of memory | $0.028 per GB of GPU memory per hour | $0.015 per GB of GPU memory per hour | H200 1.91× higher |
| Fresh listings | 18 from 11 providers | 2 from 1 providers | — |
Prices cover every version and instance size the providers list, so they are not tied to the versions compared above; the price per GB of memory divides the lowest fresh on-demand price by the memory of the compared version. Listed prices are not like for like across providers. See every listing: H200 prices · MI325X prices.
More on these GPUs
Cite this page
Last updated Oct 25, 2026. Figures may change as records are updated — include the date you retrieved them.
| APA | Data Center Census. (2026). H200 vs MI325X. Retrieved from https://datacentercensus.com/gpu-prices/compare/h200-vs-mi325x |
|---|---|
| MLA | “H200 vs MI325X.” Data Center Census, Oct 25, 2026, https://datacentercensus.com/gpu-prices/compare/h200-vs-mi325x. |
| Link | https://datacentercensus.com/gpu-prices/compare/h200-vs-mi325x |
Republishing a figure or table? Attribute it to Data Center Census with a link to this page. The underlying data is downloadable as JSON and CSV; the press kit has standing figures and contact details.