DeepSeek R1 — GB200 NVL72 vs MI300X
Head-to-head AI inference benchmark comparison of GB200 NVL72 (NVIDIA Blackwell) and MI300X (AMD CDNA 3) 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.
GB200 NVL72 / MI300X on DeepSeek R1 at 37 tok/s/user: 7073 / 513 tok/s/chip, $0.07 / $0.51 per million tokens. GB200 NVL72 is 605% cheaper per token; GB200 NVL72 delivers 1280% more tok/s/chip.
Around the middle of the 32–51 tok/s/user interactivity band, at 42 tok/s/user on DeepSeek R1: GB200 NVL72 runs 7000 tok/s/chip at $0.07/M tokens, MI300X runs 406 at $0.65/M. GB200 NVL72 is 780% cheaper per token; GB200 NVL72 delivers 1623% more tok/s/chip.
Setting 46 tok/s/user as the target on DeepSeek R1, GB200 NVL72 produces 6930 tok/s/chip ($0.07 per million tokens) and MI300X produces 322 ($0.82). GB200 NVL72 is 998% cheaper per token; GB200 NVL72 delivers 2050% more tok/s/chip. (Numbers reflect the default 8k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | GB200 NVL72:7072.6MI300X:512.6 | GB200 NVL72:7000.5MI300X:406.3 | GB200 NVL72:6930.5MI300X:322.4 |
| Cost ($/M tok) | GB200 NVL72:$0.073MI300X:$0.515 | GB200 NVL72:$0.074MI300X:$0.650 | GB200 NVL72:$0.075MI300X:$0.819 |
| tok/s/MW | GB200 NVL72:3782124MI300X:368760 | GB200 NVL72:3743569MI300X:292288 | GB200 NVL72:3706145MI300X:231934 |
| Concurrency | GB200 NVL72:~2484MI300X:~14 | GB200 NVL72:~4461MI300X:~10 | GB200 NVL72:~5881MI300X:~7 |
Inference Performance
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.
Total Tokens per $1 USD (Owning - Hyperscaler) vs. Interactivity
DeepSeek R1 0528 671B • FP4 • 8K / 1K • Source: SemiAnalysis InferenceX™
TCO $/chip/hr: H100: 1.17H200: 1.22B200: 1.73B300: 2.26GB200: 1.86GB300: 2.31MI300X: 0.95MI325X: 1.1MI355X: 1.5RTX6000PRO: 0.68
Source: SemiAnalysis Market July 2026 Pricing Surveys & AI Cloud TCO Model
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate cost per million tokens per decode chip or per prefill chip, rather than per total chip count. This makes direct token cost comparison with aggregated configs not an apples-to-apples comparison.
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate input throughput per decode chip or per prefill chip, rather than per total chip count. This makes direct input throughput comparison with aggregated configs not an apples-to-apples comparison.
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate output throughput per decode chip or per prefill chip, rather than per total chip count. This makes direct output throughput comparison with aggregated configs not an apples-to-apples comparison.
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate power per decode chip or per prefill chip, rather than per total chip count. This makes direct power comparison with aggregated configs not an apples-to-apples comparison.
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate Joules per decode chip or per prefill chip, rather than per total chip count. This makes direct Joules per token comparison with aggregated configs not an apples-to-apples comparison.
No data available
Please change the model, sequence, precision, date range or chip selection.
Shift+Scroll to zoom • Drag to pan • Double-click to reset • Click a point to pin tooltip