Head-to-head AI inference benchmark comparison of H200 (NVIDIA Hopper) and MI325X (AMD CDNA 3) on gpt-oss 120B. 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.
H200 posts 10029 tok/s/chip for $0.03 per million tokens at 44 tok/s/user on gpt-oss 120B; MI325X posts 5346 tok/s/chip for $0.06. H200 is 69% cheaper per token; H200 delivers 88% more tok/s/chip.
Throughput at 65 tok/s/user on gpt-oss 120B: H200 hits 8019 tok/s/chip, MI325X hits 3875. Per-million costs land at $0.04 and $0.08 respectively. H200 is 87% cheaper per token; H200 delivers 107% more tok/s/chip.
H200 / MI325X on gpt-oss 120B at 86 tok/s/user: 6958 / 1856 tok/s/chip, $0.05 / $0.16 per million tokens. H200 is 238% cheaper per token; H200 delivers 275% more tok/s/chip. (Numbers reflect the default 8k/1k · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | H200:10028.5MI325X:5345.8 | H200:8019.3MI325X:3874.9 | H200:6957.7MI325X:1855.5 |
| Cost ($/M tok) | H200:$0.034MI325X:$0.057 | H200:$0.042MI325X:$0.079 | H200:$0.049MI325X:$0.165 |
| tok/s/MW | H200:7320076MI325X:3163206 | H200:5853481MI325X:2292832 | H200:5078638MI325X:1097950 |
| Concurrency | H200:~36MI325X:~14 | H200:~62MI325X:~8 | H200:~25MI325X:~14 |
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.
Matching measurements exist, but their chip series are hidden.
Shift+Scroll to zoom • Drag to pan • Double-click to reset • Click a point to pin tooltip
gpt-oss 120B 8K / 1K Total Tokens per $1 TCO vs. Interactivity