MiniMax M3 428B — B200 vs GB200 NVL72
Head-to-head AI inference benchmark comparison of B200 (NVIDIA Blackwell) and GB200 NVL72 (NVIDIA Blackwell) on MiniMax M3 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 →
Near the low end of the 44–242 tok/s/user interactivity band, at 93 tok/s/user on MiniMax M3 428B: B200 runs 46184 tok/s/chip at $0.01/M tokens, GB200 NVL72 runs 29608 at $0.02/M. B200 is 68% cheaper per token; B200 delivers 56% more tok/s/chip.
Setting 143 tok/s/user as the target on MiniMax M3 428B, B200 produces 38427 tok/s/chip ($0.01 per million tokens) and GB200 NVL72 produces 20231 ($0.03). B200 is 104% cheaper per token; B200 delivers 90% more tok/s/chip.
At 193 tok/s/user interactivity on MiniMax M3 428B, B200 delivers 26013 tok/s/chip at $0.02 per million tokens; GB200 NVL72 delivers 8390 tok/s/chip at $0.06. B200 is 233% cheaper per token; B200 delivers 210% more tok/s/chip at this point. (Numbers reflect the default agentic-traces · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B200:46184.0GB200 NVL72:29608.2 | B200:38427.0GB200 NVL72:20231.1 | B200:26013.1GB200 NVL72:8389.8 |
| Cost ($/M tok) | B200:$0.010GB200 NVL72:$0.017 | B200:$0.013GB200 NVL72:$0.026 | B200:$0.018GB200 NVL72:$0.062 |
| tok/s/MW | B200:27008179GB200 NVL72:15833287 | B200:22471956GB200 NVL72:10818785 | B200:15212346GB200 NVL72:4486537 |
| Concurrency | B200:~32GB200 NVL72:~20 | B200:~23GB200 NVL72:~14 | B200:~17GB200 NVL72:~5 |
Total Tokens per $1 USD (Owning - Neocloud Giant) vs. Interactivity
MiniMax M3 428B • FP4 • 8K / 1K • Source: SemiAnalysis InferenceX™
MiniMax M3 428B • FP4 • 8K / 1K • Source: SemiAnalysis InferenceX™TCO $/chip/hr: VR200: 3.61H100: 1.55H200: 1.59B200: 2.07B300: 2.52GB200: 2.26GB300: 2.79MI300X: 1.16MI325X: 1.32MI355X: 2.09RTX6000PRO: 0.75Jalapeño (Teacup): 1.56
Source: SemiAnalysis Market July 2026 Pricing Surveys & AI Cloud TCO Model
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate cost per million tokens per decode chip or per prefill chip, rather than per total chip count. This makes direct token cost comparison with aggregated configs not an apples-to-apples comparison.
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate input throughput per decode chip or per prefill chip, rather than per total chip count. This makes direct input throughput comparison with aggregated configs not an apples-to-apples comparison.
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate output throughput per decode chip or per prefill chip, rather than per total chip count. This makes direct output throughput comparison with aggregated configs not an apples-to-apples comparison.
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate power per decode chip or per prefill chip, rather than per total chip count. This makes direct power comparison with aggregated configs not an apples-to-apples comparison.
Note: Disaggregated inference configurations (e.g., MoRI SGLang, Dynamo TRTLLM) calculate Joules per decode chip or per prefill chip, rather than per total chip count. This makes direct Joules per token comparison with aggregated configs not an apples-to-apples comparison.
No data available
Please change the model, sequence, precision, date range or chip selection.
Shift+Scroll to zoom • Drag to pan • Double-click to reset • Click a point to pin tooltip
Interactivity is the rate at which a single user receives generated tokens while the model streams its answer — how quickly new words appear on screen. Higher values feel snappier; operators trade it against batch throughput.
How many total tokens (input + output) one US dollar of infrastructure spend buys, priced with the all-in hourly ownership cost of a Neocloud Giant operator. It is the reciprocal of cost per token, so higher means cheaper.
Formula: tok/$ = (total tok/s/chip × 3,600) ÷ all-in cost per chip-hour ($)