DeepSeek R1 — GB300 NVL72 vs MI355X
Head-to-head AI inference benchmark comparison of GB300 NVL72 (NVIDIA Blackwell) and MI355X (AMD CDNA 4) on DeepSeek R1. 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.
Throughput at 66 tok/s/user on DeepSeek R1: GB300 NVL72 hits 3550 tok/s/chip, MI355X hits 807. Per-million costs land at $0.18 and $0.50 respectively. GB300 NVL72 is 177% cheaper per token; GB300 NVL72 delivers 340% more tok/s/chip.
GB300 NVL72 / MI355X on DeepSeek R1 at 116 tok/s/user: 1300 / 295 tok/s/chip, $0.48 / $1.43 per million tokens. GB300 NVL72 is 194% cheaper per token; GB300 NVL72 delivers 340% more tok/s/chip.
Toward the upper edge of the 17–216 tok/s/user interactivity band, at 166 tok/s/user on DeepSeek R1: GB300 NVL72 runs 177 tok/s/chip at $3.73/M tokens, MI355X runs 130 at $3.64/M. MI355X is 3% cheaper per token; GB300 NVL72 delivers 36% more tok/s/chip. (Numbers reflect the default 1k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | GB300 NVL72:3549.8MI355X:807.1 | GB300 NVL72:1299.6MI355X:295.1 | GB300 NVL72:177.2MI355X:130.1 |
| Cost ($/M tok) | GB300 NVL72:$0.181MI355X:$0.503 | GB300 NVL72:$0.484MI355X:$1.426 | GB300 NVL72:$3.734MI355X:$3.637 |
| tok/s/MW | GB300 NVL72:1674439MI355X:386158 | GB300 NVL72:613011MI355X:141218 | GB300 NVL72:83566MI355X:62242 |
| Concurrency | GB300 NVL72:~1151MI355X:~115 | GB300 NVL72:~337MI355X:~10 | GB300 NVL72:~32MI355X:~4 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.