Llama 3.3 70B — H100 vs H200
Head-to-head AI inference benchmark comparison of H100 (NVIDIA Hopper) and H200 (NVIDIA Hopper) on Llama 3.3 70B. 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 38 tok/s/user on Llama 3.3 70B: H100 hits 1848 tok/s/chip, H200 hits 3310. Per-million costs land at $0.18 and $0.10 respectively. H200 is 72% cheaper per token; H200 delivers 79% more tok/s/chip.
H100 / H200 on Llama 3.3 70B at 58 tok/s/user: 1171 / 2516 tok/s/chip, $0.28 / $0.13 per million tokens. H200 is 106% cheaper per token; H200 delivers 115% more tok/s/chip.
Toward the upper edge of the 19–97 tok/s/user interactivity band, at 78 tok/s/user on Llama 3.3 70B: H100 runs 877 tok/s/chip at $0.37/M tokens, H200 runs 1864 at $0.18/M. H200 is 104% cheaper per token; H200 delivers 112% more tok/s/chip. (Numbers reflect the default 8k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | H100:1847.6H200:3310.0 | H100:1171.4H200:2516.2 | H100:877.4H200:1864.0 |
| Cost ($/M tok) | H100:$0.176H200:$0.102 | H100:$0.277H200:$0.135 | H100:$0.370H200:$0.182 |
| tok/s/MW | H100:1348638H200:2416052 | H100:855073H200:1836618 | H100:640410H200:1360591 |
| Concurrency | H100:~20H200:~32 | H100:~16H200:~23 | H100:~8H200:~10 |
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
Llama 3.3 70B Instruct • 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