Head-to-head AI inference benchmark comparison of H100 (NVIDIA Hopper) 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.
At 28 tok/s/user interactivity on DeepSeek R1, H100 delivers 850 tok/s/chip at $0.38 per million tokens; MI300X delivers 656 tok/s/chip at $0.40. H100 is 5% cheaper per token; H100 delivers 29% more tok/s/chip at this point.
H100 posts 850 tok/s/chip for $0.38 per million tokens at 36 tok/s/user on DeepSeek R1; MI300X posts 532 tok/s/chip for $0.50. H100 is 30% cheaper per token; H100 delivers 60% more tok/s/chip.
Throughput at 44 tok/s/user on DeepSeek R1: H100 hits 761 tok/s/chip, MI300X hits 364. Per-million costs land at $0.43 and $0.73 respectively. H100 is 70% cheaper per token; H100 delivers 109% more tok/s/chip. (Numbers reflect the default 8k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | H100:850.1MI300X:656.4 | H100:850.1MI300X:532.4 | H100:760.9MI300X:363.7 |
| Cost ($/M tok) | H100:$0.382MI300X:$0.402 | H100:$0.382MI300X:$0.496 | H100:$0.427MI300X:$0.726 |
| tok/s/MW | H100:620476MI300X:472258 | H100:620476MI300X:383002 | H100:555426MI300X:261647 |
| Concurrency | H100:~154MI300X:~22 | H100:~154MI300X:~15 | H100:~136MI300X:~8 |
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.
TCO $/chip/hr: VR200: 3.61H100: 1.17H200: 1.22B200: 1.73B300: 2.26GB200: 1.86GB300: 2.31MI300X: 0.95MI325X: 1.1MI355X: 1.5RTX6000PRO: 0.68Jalapeño (Teacup): 1.47
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 measurements to plot for this selection. Review the benchmark controls above or adjust quick filters.
Shift+Scroll to zoom • Drag to pan • Double-click to reset • Click a point to pin tooltip
DeepSeek R1 0528 671B 8K / 1K Total Tokens per $1 TCO vs. Interactivity