GLM 5/5.1 — GB300 NVL72 vs MI325X
Head-to-head AI inference benchmark comparison of GB300 NVL72 (NVIDIA Blackwell) and MI325X (AMD CDNA 3) on GLM 5/5.1. 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.
MI325X: 173 tok/s/GPU, $2.08 per million tokens at 21 tok/s/user on GLM 5/5.1. GB300 NVL72 is unmeasured here.
At 26 tok/s/user on GLM 5/5.1, MI325X delivers 101 tok/s/GPU at $3.53 per million tokens; GB300 NVL72 hasn't been benchmarked at this target.
MI325X hits 62 tok/s/GPU for $5.71 per million tokens at 30 tok/s/user on GLM 5/5.1. No GB300 NVL72 data at this operating point. (Numbers reflect the default 1k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/gpu) | GB300 NVL72:—MI325X:172.9 | GB300 NVL72:—MI325X:101.3 | GB300 NVL72:—MI325X:61.6 |
| Cost ($/M tok) | GB300 NVL72:—MI325X:$2.081 | GB300 NVL72:—MI325X:$3.534 | GB300 NVL72:—MI325X:$5.710 |
| tok/s/MW | GB300 NVL72:—MI325X:79293 | GB300 NVL72:—MI325X:46466 | GB300 NVL72:—MI325X:28234 |
| Concurrency | GB300 NVL72:—MI325X:~34 | GB300 NVL72:—MI325X:~16 | GB300 NVL72:—MI325X:~9 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.