Head-to-head AI inference benchmark comparison of B300 (NVIDIA Blackwell) and GB200 NVL72 (NVIDIA Blackwell) 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
Near the low end of the 10β549 tok/s/user interactivity band, at 145 tok/s/user on MiniMax M3.1 Flash 428B: B300 runs 61216 tok/s/chip at $0.01/M tokens, GB200 NVL72 runs 51701 at $0.01/M. GB200 NVL72 is 3% cheaper per token; B300 delivers 18% more tok/s/chip.
Setting 280 tok/s/user as the target on MiniMax M3.1 Flash 428B, B300 produces 22779 tok/s/chip ($0.03 per million tokens) and GB200 NVL72 produces 33217 ($0.02). GB200 NVL72 is 77% cheaper per token; GB200 NVL72 delivers 46% more tok/s/chip.
At 414 tok/s/user interactivity on MiniMax M3.1 Flash 428B, B300 delivers 9547 tok/s/chip at $0.07 per million tokens; GB200 NVL72 delivers 15524 tok/s/chip at $0.03. GB200 NVL72 is 98% cheaper per token; GB200 NVL72 delivers 63% more tok/s/chip at this point. (Numbers reflect the default agentic-traces Β· fp4 selection for this URL β table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B300:61215.8GB200 NVL72:51701.4 | B300:22778.5GB200 NVL72:33216.6 | B300:9546.8GB200 NVL72:15523.8 |
| Cost ($/M tok) | B300:$0.010GB200 NVL72:$0.010 | B300:$0.028GB200 NVL72:$0.016 | B300:$0.066GB200 NVL72:$0.033 |
| tok/s/MW | B300:32218854GB200 NVL72:27647785 | B300:11988709GB200 NVL72:17762895 | B300:5024651GB200 NVL72:8301522 |
| Concurrency | B300:~40GB200 NVL72:~32 | B300:~12GB200 NVL72:~19 | B300:~4GB200 NVL72:~8 |
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.
