H200 SXM vs MI325X

Open Advisor →

Side-by-side GPU comparison: specs, memory, compute performance, and live cloud pricing.

Verdict

MI325X has 115 GB more VRAM, making it better suited for large models and long context windows. For compute-bound workloads like training, MI325X delivers 1.3× higher FP16 throughput. MI325X supports a broader range of models (671 vs 657 from this catalog), giving more flexibility.

Specifications

H200 SXMMI325X
VRAM141 GB256 GB
VRAM TypeHBM3eHBM3e
Memory Bandwidth4.8 TB/s6.0 TB/s
FP16 Performance990 TFLOPS1307 TFLOPS
ManufacturerNVIDIAAMD
FP8 SupportYesYes
FP4 SupportNoNo

Price / Performance

Based on cheapest single-GPU on-demand pricing. Lower $/TFLOP = better compute value; lower $/GB = better memory value.

H200 SXMMI325X
$/hr (cheapest)$2.35
$/TFLOP (compute value)$0.0024
$/GB VRAM (memory value)$0.0167

Cloud Pricing

Cheapest on-demand price per provider (single GPU).

H200 SXM

ProviderOn-demandSpotRent
Vast.ai$2.35/hrRent
Nebius$3.50/hr$1.45/hr

MI325X

No cloud pricing available.

Model Compatibility

Models from the catalog that fit on each GPU, grouped by required precision.

H200 SXM (657 models)

MI325X (671 models)

You might also compare…

Pricing data refreshed hourly · Last updated April 11, 2026 · Browse all comparisons