Head-to-head AI inference benchmark comparison of B300 (NVIDIA Blackwell) 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 →
Setting 75 tok/s/user as the target on DeepSeek V4.1 Flash 552B, B300 produces 76422 tok/s/chip ($0.01 per million tokens) and H200 produces 11007 ($0.03). B300 is 275% cheaper per token; B300 delivers 594% more tok/s/chip.
At 124 tok/s/user interactivity on DeepSeek V4.1 Flash 552B, B300 delivers 46437 tok/s/chip at $0.01 per million tokens; H200 delivers 6938 tok/s/chip at $0.05. B300 is 261% cheaper per token; B300 delivers 569% more tok/s/chip at this point.
B300 posts 28520 tok/s/chip for $0.02 per million tokens at 173 tok/s/user on DeepSeek V4.1 Flash 552B; H200 posts 4778 tok/s/chip for $0.07. B300 is 222% cheaper per token; B300 delivers 497% more tok/s/chip. (Numbers reflect the default agentic-traces · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B300:76421.7H200:11006.9 | B300:46437.2H200:6938.1 | B300:28520.3H200:4777.7 |
| Cost ($/M tok) | B300:$0.008H200:$0.031 | B300:$0.014H200:$0.049 | B300:$0.022H200:$0.071 |
| tok/s/MW | B300:40221969H200:8034217 | B300:24440629H200:5064304 | B300:15010703H200:3487364 |
| Concurrency | B300:~46H200:~15 | B300:~25H200:~8 | B300:~16H200:~5 |
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.
