GPU comparison

H100 vs MI300X

Head-to-head AI inference benchmark comparison of H100 (NVIDIA Hopper) and MI300X (AMD CDNA 3). 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)
H100:298.8MI300X:176.4
H100:287.3MI300X:142.3
H100:279.4MI300X:97.2
Cost ($/M tok)
H100:$1.216MI300X:$1.752
H100:$1.269MI300X:$2.115
H100:$1.305MI300X:$3.201
tok/s/MW
H100:172740MI300X:98566
H100:166067MI300X:79513
H100:161495MI300X:54314
Concurrency
H100:~828MI300X:~30
H100:~732MI300X:~20
H100:~666MI300X:~11

Inference Performance

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