Head-to-head AI inference benchmark comparison of H100 (NVIDIA Hopper) and H200 (NVIDIA Hopper) on DeepSeek V4.1 Flash 552B. 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.
AgentX replays real coding-agent sessions rather than fixed-length prompts, so context grows turn over turn and most of each request is served from cache instead of being recomputed. That turns the comparison into a systems question: KV transfer between nodes, prefix-aware routing, and cache capacity all move the curve alongside raw chip throughput. Fixed-sequence workloads stay the clean baseline for kernel and silicon performance, so the two scenarios answer different questions about the same hardware. Learn more about AgentX →
H100 / H200 on DeepSeek V4.1 Flash 552B at 81 tok/s/user: 10947 / 19845 tok/s/chip, $0.03 / $0.02 per million tokens. H200 is 74% cheaper per token; H200 delivers 81% more tok/s/chip.
Around the middle of the 7–303 tok/s/user interactivity band, at 155 tok/s/user on DeepSeek V4.1 Flash 552B: H100 runs 6922 tok/s/chip at $0.05/M tokens, H200 runs 11019 at $0.03/M. H200 is 53% cheaper per token; H200 delivers 59% more tok/s/chip.
Setting 230 tok/s/user as the target on DeepSeek V4.1 Flash 552B, H100 produces 4187 tok/s/chip ($0.08 per million tokens) and H200 produces 5740 ($0.06). H200 is 31% cheaper per token; H200 delivers 37% more tok/s/chip. (Numbers reflect the default agentic-traces · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | H100:10946.7H200:19845.0 | H100:6921.5H200:11018.6 | H100:4187.3H200:5740.5 |
| Cost ($/M tok) | H100:$0.030H200:$0.017 | H100:$0.047H200:$0.031 | H100:$0.078H200:$0.059 |
| tok/s/MW | H100:7990287H200:14485434 | H100:5052199H200:8042768 | H100:3056440H200:4190122 |
| Concurrency | H100:~15H200:~14 | H100:~8H200:~6 | H100:~4H200:~2 |
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.
No measurements to plot for this selection. Review the benchmark controls above or adjust quick filters.
