GLM 5.2 · Chip comparison
GLM 5.2 — H200 vs MI355X
Head-to-head AI inference benchmark comparison of H200 (NVIDIA Hopper) and MI355X (AMD CDNA 4) on GLM 5.2. Latency, throughput, and cost across LLM workloads. Use the chart controls below to switch sequences, precisions, and metrics — same interactions as the main inference chart.
No interpolated comparison data available for the default model. Use the chart controls below to select a model with benchmark data for both chips.
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.
Vendor:
Deployment:
Spec Decoding: