GPU comparison

H200 vs MI355X

Head-to-head AI inference benchmark comparison of H200 (NVIDIA Hopper) and MI355X (AMD CDNA 4). Latency, throughput, and cost across LLM workloads. Use the chart controls below to switch models, sequences, precisions, and metrics — same interactions as the main inference chart.

Interpolated from real benchmark data. Edit target interactivity values below to compare at different operating points.
Metric
Interactivity (tok/s/user)
Interactivity (tok/s/user)
Interactivity (tok/s/user)
Throughput (tok/s/gpu)
H200:539.5MI355X:902.7
H200:138.0MI355X:309.1
H200:39.6MI355X:174.8
Cost ($/M tok)
H200:$0.751MI355X:$0.457
H200:$2.877MI355X:$1.345
H200:$9.940MI355X:$2.420
tok/s/MW
H200:311843MI355X:340650
H200:79785MI355X:116634
H200:22869MI355X:65969
Concurrency
H200:~390MI355X:~61
H200:~67MI355X:~23
H200:~14MI355X:~4

Inference Performance

Inference performance metrics across different models, hardware configurations, and serving parameters.