Head-to-head AI inference benchmark comparison of B200 (NVIDIA Blackwell) and H200 (NVIDIA Hopper) on MiniMax M3.1 Flash 428B. 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
Setting 87 tok/s/user as the target on MiniMax M3.1 Flash 428B, B200 produces 58986 tok/s/chip ($0.01 per million tokens) and H200 produces 9696 ($0.03). B200 is 329% cheaper per token; B200 delivers 508% more tok/s/chip.
At 156 tok/s/user interactivity on MiniMax M3.1 Flash 428B, B200 delivers 58986 tok/s/chip at $0.01 per million tokens; H200 delivers 6955 tok/s/chip at $0.05. B200 is 498% cheaper per token; B200 delivers 748% more tok/s/chip at this point.
B200 posts 35737 tok/s/chip for $0.01 per million tokens at 225 tok/s/user on MiniMax M3.1 Flash 428B; H200 posts 4152 tok/s/chip for $0.08. B200 is 507% cheaper per token; B200 delivers 761% more tok/s/chip. (Numbers reflect the default agentic-traces Β· fp4 selection for this URL β table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B200:58986.4H200:9696.3 | B200:58986.4H200:6954.5 | B200:35737.3H200:4151.7 |
| Cost ($/M tok) | B200:$0.008H200:$0.035 | B200:$0.008H200:$0.049 | B200:$0.013H200:$0.082 |
| tok/s/MW | B200:34494965H200:7077580 | B200:34494965H200:5076287 | B200:20899029H200:3030419 |
| Concurrency | B200:~32H200:~15 | B200:~32H200:~9 | B200:~19H200:~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.
