DeepSeek V4 Pro 1.6T — B300 vs MI355X
Head-to-head AI inference benchmark comparison of B300 (NVIDIA Blackwell) and MI355X (AMD CDNA 4) 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.
B300 posts 5451 tok/s/GPU for $0.12 per million tokens at 32 tok/s/user on DeepSeek V4 Pro 1.6T; MI355X posts 2861 tok/s/GPU for $0.14. B300 is 21% cheaper per token; B300 delivers 91% more tok/s/GPU.
Throughput at 62 tok/s/user on DeepSeek V4 Pro 1.6T: B300 hits 3257 tok/s/GPU, MI355X hits 1070. Per-million costs land at $0.21 and $0.38 respectively. B300 is 84% cheaper per token; B300 delivers 204% more tok/s/GPU.
B300 / MI355X on DeepSeek V4 Pro 1.6T at 91 tok/s/user: 1806 / 601 tok/s/GPU, $0.35 / $0.70 per million tokens. B300 is 97% cheaper per token; B300 delivers 200% more tok/s/GPU. (Numbers reflect the default 8k/1k · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/gpu) | B300:5451.5MI355X:2860.6 | B300:3257.1MI355X:1069.9 | B300:1805.9MI355X:601.3 |
| Cost ($/M tok) | B300:$0.120MI355X:$0.145 | B300:$0.208MI355X:$0.384 | B300:$0.354MI355X:$0.699 |
| tok/s/MW | B300:2869198MI355X:1368719 | B300:1714265MI355X:511918 | B300:950485MI355X:287699 |
| Concurrency | B300:~1024MI355X:~101 | B300:~76MI355X:~16 | B300:~11MI355X:~7 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.