Kimi K2.5/K2.6/K2.7-Code 1T — GB300 NVL72 vs MI300X
Head-to-head AI inference benchmark comparison of GB300 NVL72 (NVIDIA Blackwell) and MI300X (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.
At 60 tok/s/user on Kimi K2.5/K2.6/K2.7-Code 1T, GB300 NVL72 delivers 7910 tok/s/GPU at $0.09 per million tokens; MI300X hasn't been benchmarked at this target.
GB300 NVL72 hits 576 tok/s/GPU for $1.28 per million tokens at 104 tok/s/user on Kimi K2.5/K2.6/K2.7-Code 1T. No MI300X data at this operating point.
GB300 NVL72: 194 tok/s/GPU, $3.76 per million tokens at 148 tok/s/user on Kimi K2.5/K2.6/K2.7-Code 1T. MI300X is unmeasured here. (Numbers reflect the default 1k/1k · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/gpu) | GB300 NVL72:7910.3MI300X:— | GB300 NVL72:576.5MI300X:— | GB300 NVL72:194.0MI300X:— |
| Cost ($/M tok) | GB300 NVL72:$0.091MI300X:— | GB300 NVL72:$1.285MI300X:— | GB300 NVL72:$3.759MI300X:— |
| tok/s/MW | GB300 NVL72:3731285MI300X:— | GB300 NVL72:271928MI300X:— | GB300 NVL72:91527MI300X:— |
| Concurrency | GB300 NVL72:~3377MI300X:— | GB300 NVL72:~119MI300X:— | GB300 NVL72:~28MI300X:— |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.