Head-to-head AI inference benchmark comparison of B200 (NVIDIA Blackwell) and H100 (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 β 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 β
B200 / H100 on DeepSeek V4.1 Flash 552B at 125 tok/s/user: 309889 / 15080 tok/s/chip, $0.00 / $0.02 per million tokens. B200 is 1290% cheaper per token; B200 delivers 1955% more tok/s/chip.
Around the middle of the 60β322 tok/s/user interactivity band, at 191 tok/s/user on DeepSeek V4.1 Flash 552B: B200 runs 163039 tok/s/chip at $0.00/M tokens, H100 runs 6745 at $0.05/M. B200 is 1535% cheaper per token; B200 delivers 2317% more tok/s/chip.
Setting 257 tok/s/user as the target on DeepSeek V4.1 Flash 552B, B200 produces 102789 tok/s/chip ($0.00 per million tokens) and H100 produces 4302 ($0.08). B200 is 1516% cheaper per token; B200 delivers 2289% more tok/s/chip. (Numbers reflect the default agentic-traces Β· fp4 selection for this URL β table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B200:309889.2H100:15080.3 | B200:163038.9H100:6744.6 | B200:102788.7H100:4302.1 |
| Cost ($/M tok) | B200:$0.002H100:$0.022 | B200:$0.003H100:$0.048 | B200:$0.005H100:$0.076 |
| tok/s/MW | B200:181221768H100:11007505 | B200:95344388H100:4923070 | B200:60110337H100:3140225 |
| Concurrency | B200:~218H100:~18 | B200:~102H100:~7 | B200:~55H100:~4 |
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.
