Head-to-head AI inference benchmark comparison of GB200 NVL72 (NVIDIA Blackwell) and GB300 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 10β529 tok/s/user interactivity band, at 140 tok/s/user on MiniMax M3.1 Flash 428B: GB200 NVL72 runs 46476 tok/s/chip at $0.01/M tokens, GB300 NVL72 runs 57954 at $0.01/M. Cost per token is essentially tied; GB300 NVL72 delivers 25% more tok/s/chip.
Setting 269 tok/s/user as the target on MiniMax M3.1 Flash 428B, GB200 NVL72 produces 21804 tok/s/chip ($0.02 per million tokens) and GB300 NVL72 produces 22405 ($0.03). GB200 NVL72 is 21% cheaper per token; GB300 NVL72 delivers 3% more tok/s/chip.
At 399 tok/s/user interactivity on MiniMax M3.1 Flash 428B, GB200 NVL72 delivers 11598 tok/s/chip at $0.04 per million tokens; GB300 NVL72 delivers 12583 tok/s/chip at $0.05. GB200 NVL72 is 14% cheaper per token; GB300 NVL72 delivers 8% more tok/s/chip at this point. (Numbers reflect the default agentic-traces Β· fp4 selection for this URL β table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | GB200 NVL72:46476.0GB300 NVL72:57953.6 | GB200 NVL72:21803.8GB300 NVL72:22404.7 | GB200 NVL72:11598.2GB300 NVL72:12582.7 |
| Cost ($/M tok) | GB200 NVL72:$0.011GB300 NVL72:$0.011 | GB200 NVL72:$0.024GB300 NVL72:$0.029 | GB200 NVL72:$0.045GB300 NVL72:$0.051 |
| tok/s/MW | GB200 NVL72:24853478GB300 NVL72:27336609 | GB200 NVL72:11659760GB300 NVL72:10568270 | GB200 NVL72:6202255GB300 NVL72:5935231 |
| Concurrency | GB200 NVL72:~28GB300 NVL72:~32 | GB200 NVL72:~12GB300 NVL72:~12 | GB200 NVL72:~6GB300 NVL72:~6 |
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.
