Head-to-head AI inference benchmark comparison of GB200 NVL72 (NVIDIA Blackwell) and H100 (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
Throughput at 84 tok/s/user on MiniMax M3.1 Flash 428B: GB200 NVL72 hits 52698 tok/s/chip, H100 hits 5732. Per-million costs land at $0.01 and $0.06 respectively. GB200 NVL72 is 478% cheaper per token; GB200 NVL72 delivers 819% more tok/s/chip.
GB200 NVL72 / H100 on MiniMax M3.1 Flash 428B at 152 tok/s/user: 43465 / 4486 tok/s/chip, $0.01 / $0.07 per million tokens. GB200 NVL72 is 509% cheaper per token; GB200 NVL72 delivers 869% more tok/s/chip.
Toward the upper edge of the 17β288 tok/s/user interactivity band, at 220 tok/s/user on MiniMax M3.1 Flash 428B: GB200 NVL72 runs 28671 tok/s/chip at $0.02/M tokens, H100 runs 3595 at $0.09/M. GB200 NVL72 is 402% cheaper per token; GB200 NVL72 delivers 698% more tok/s/chip. (Numbers reflect the default agentic-traces Β· fp4 selection for this URL β table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | GB200 NVL72:52698.4H100:5732.4 | GB200 NVL72:43464.7H100:4486.1 | GB200 NVL72:28670.6H100:3594.6 |
| Cost ($/M tok) | GB200 NVL72:$0.010H100:$0.057 | GB200 NVL72:$0.012H100:$0.072 | GB200 NVL72:$0.018H100:$0.090 |
| tok/s/MW | GB200 NVL72:28180958H100:4184237 | GB200 NVL72:23243173H100:3274527 | GB200 NVL72:15331851H100:2623785 |
| Concurrency | GB200 NVL72:~32H100:~9 | GB200 NVL72:~26H100:~7 | GB200 NVL72:~16H100:~3 |
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.
