DeepSeek V4 Pro 1.6T — B200 vs MI325X
Head-to-head AI inference benchmark comparison of B200 (NVIDIA Blackwell) and MI325X (AMD CDNA 3) on DeepSeek V4 Pro 1.6T. 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.
B200: 3770 tok/s/GPU, $0.14 per million tokens at 61 tok/s/user on DeepSeek V4 Pro 1.6T. MI325X is unmeasured here.
At 116 tok/s/user on DeepSeek V4 Pro 1.6T, B200 delivers 882 tok/s/GPU at $0.61 per million tokens; MI325X hasn't been benchmarked at this target.
B200 hits 485 tok/s/GPU for $1.14 per million tokens at 171 tok/s/user on DeepSeek V4 Pro 1.6T. No MI325X data at this operating point. (Numbers reflect the default 8k/1k · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/gpu) | B200:3770.5MI325X:— | B200:882.2MI325X:— | B200:485.0MI325X:— |
| Cost ($/M tok) | B200:$0.144MI325X:— | B200:$0.613MI325X:— | B200:$1.144MI325X:— |
| tok/s/MW | B200:2204965MI325X:— | B200:515922MI325X:— | B200:283620MI325X:— |
| Concurrency | B200:~76MI325X:— | B200:~8MI325X:— | B200:~3MI325X:— |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.