Kimi K2.5/K2.6/K2.7-Code 1T — GB200 NVL72 vs MI325X
Head-to-head AI inference benchmark comparison of GB200 NVL72 (NVIDIA Blackwell) and MI325X (AMD CDNA 3) on Kimi K2.5/K2.6/K2.7-Code 1T. 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.
GB200 NVL72 hits 7627 tok/s/GPU for $0.08 per million tokens at 60 tok/s/user on Kimi K2.5/K2.6/K2.7-Code 1T. No MI325X data at this operating point.
GB200 NVL72: 451 tok/s/GPU, $1.41 per million tokens at 102 tok/s/user on Kimi K2.5/K2.6/K2.7-Code 1T. MI325X is unmeasured here.
At 144 tok/s/user on Kimi K2.5/K2.6/K2.7-Code 1T, GB200 NVL72 delivers 192 tok/s/GPU at $3.15 per million tokens; MI325X hasn't been benchmarked at this target. (Numbers reflect the default 1k/1k · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/gpu) | GB200 NVL72:7626.9MI325X:— | GB200 NVL72:451.0MI325X:— | GB200 NVL72:192.1MI325X:— |
| Cost ($/M tok) | GB200 NVL72:$0.080MI325X:— | GB200 NVL72:$1.407MI325X:— | GB200 NVL72:$3.151MI325X:— |
| tok/s/MW | GB200 NVL72:3631859MI325X:— | GB200 NVL72:214751MI325X:— | GB200 NVL72:91476MI325X:— |
| Concurrency | GB200 NVL72:~3340MI325X:— | GB200 NVL72:~98MI325X:— | GB200 NVL72:~28MI325X:— |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.