DeepSeek R1 · Chip comparison
DeepSeek R1 — H100 vs H200
Head-to-head AI inference benchmark comparison of H100 (NVIDIA Hopper) and H200 (NVIDIA Hopper) 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.
H100 / H200 on DeepSeek R1 at 45 tok/s/user: 739 / 1,514 tok/s/chip, $0.44 / $0.22 per million tokens. H200 is 49% cheaper per token; H200 delivers 105% more tok/s/chip.
Around the middle of the 20–122 tok/s/user interactivity band, at 71 tok/s/user on DeepSeek R1: H100 runs 281 tok/s/chip at $1.16 per million tokens, H200 runs 961 at $0.35. H200 is 70% cheaper per token; H200 delivers 3.4× the tok/s/chip.
Setting 97 tok/s/user as the target on DeepSeek R1, H100 produces 154 tok/s/chip ($2.11 per million tokens) and H200 produces 726 ($0.47). H200 is 78% cheaper per token; H200 delivers 4.7× the tok/s/chip. (Numbers reflect the default 8k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
Interpolated from real benchmark data. Edit target interactivity values below to compare at different operating points.
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | H100739H2001,514 | H100281H200961 | H100154H200726 |
| Cost ($/M tok) | H100$0.44H200$0.22 | H100$1.16H200$0.35 | H100$2.11H200$0.47 |
| tok/s/MW | H100539,778H2001,105,266 | H100204,852H200701,279 | H100112,644H200529,914 |
| Concurrency | H100~132H200~33 | H100~41H200~123 | H100~17H200~7 |