MiniMax M3.1 Flash 428B · Chip comparison
MiniMax M3.1 Flash 428B — GB300 NVL72 vs Vera Rubin NVL72
Head-to-head AI inference benchmark comparison of GB300 NVL72 (NVIDIA Blackwell) and Vera Rubin NVL72 (NVIDIA Vera Rubin) on MiniMax M3.1 Flash 428B. 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 446 tok/s/user as the target on MiniMax M3.1 Flash 428B, GB300 NVL72 produces 13,714 tok/s/chip ($0.047 per million tokens) and Vera Rubin NVL72 produces 151,681 ($0.007). Vera Rubin NVL72 is 86% cheaper per token; Vera Rubin NVL72 delivers 11× the tok/s/chip.
At 517 tok/s/user interactivity on MiniMax M3.1 Flash 428B, GB300 NVL72 delivers 8,993 tok/s/chip at $0.071 per million tokens; Vera Rubin NVL72 delivers 83,149 tok/s/chip at $0.012. Vera Rubin NVL72 is 83% cheaper per token; Vera Rubin NVL72 delivers 9.2× the tok/s/chip at this point.
GB300 NVL72 posts 8,625 tok/s/chip for $0.074 per million tokens at 587 tok/s/user on MiniMax M3.1 Flash 428B; Vera Rubin NVL72 posts 65,706 tok/s/chip for $0.015. Vera Rubin NVL72 is 79% cheaper per token; Vera Rubin NVL72 delivers 7.6× the tok/s/chip. (Numbers reflect the default agentic-traces · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
Interpolated from real benchmark data. Edit target interactivity values below to compare at different operating points.
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | GB300 NVL7213,714Vera Rubin NVL72151,681 | GB300 NVL728,993Vera Rubin NVL7283,149 | GB300 NVL728,625Vera Rubin NVL7265,706 |
| Cost ($/M tok) | GB300 NVL72$0.047Vera Rubin NVL72$0.007 | GB300 NVL72$0.071Vera Rubin NVL72$0.012 | GB300 NVL72$0.074Vera Rubin NVL72$0.015 |
| tok/s/MW | GB300 NVL726,468,741Vera Rubin NVL7245,963,814 | GB300 NVL724,241,808Vera Rubin NVL7225,196,515 | GB300 NVL724,068,466Vera Rubin NVL7219,910,996 |
| Concurrency | GB300 NVL72~6Vera Rubin NVL72~200 | GB300 NVL72~4Vera Rubin NVL72~121 | GB300 NVL72~2Vera Rubin NVL72~74 |