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 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 66β432 tok/s/user interactivity band, at 157 tok/s/user on MiniMax M3 428B: B200 runs 34047 tok/s/chip at $0.01/M tokens, GB200 NVL72 runs 21155 at $0.02/M. B200 is 73% cheaper per token; B200 delivers 61% more tok/s/chip.
Setting 249 tok/s/user as the target on MiniMax M3 428B, B200 produces 17054 tok/s/chip ($0.03 per million tokens) and GB200 NVL72 produces 14144 ($0.04). B200 is 30% cheaper per token; B200 delivers 21% more tok/s/chip.
At 341 tok/s/user interactivity on MiniMax M3 428B, B200 delivers 7316 tok/s/chip at $0.07 per million tokens; GB200 NVL72 delivers 7596 tok/s/chip at $0.07. B200 is 4% cheaper per token; GB200 NVL72 delivers 4% more tok/s/chip at this point. (Numbers reflect the default agentic-traces Β· fp4 selection for this URL β table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B200:34047.1GB200 NVL72:21155.1 | B200:17054.2GB200 NVL72:14143.8 | B200:7315.6GB200 NVL72:7596.0 |
| Cost ($/M tok) | B200:$0.014GB200 NVL72:$0.024 | B200:$0.028GB200 NVL72:$0.037 | B200:$0.066GB200 NVL72:$0.068 |
| tok/s/MW | B200:19910577GB200 NVL72:11312876 | B200:9973199GB200 NVL72:7563553 | B200:4278146GB200 NVL72:4062057 |
| Concurrency | B200:~20GB200 NVL72:~14 | B200:~10GB200 NVL72:~7 | B200:~2GB200 NVL72:~1 |
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.
TCO $/chip/hr:
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.
Matching measurements exist, but their chip series are hidden.
Shift+Scroll to zoom β’ Drag to pan β’ Double-click to reset β’ Click a point to pin tooltip
MiniMax M3 428B 8K / 1K Total Tokens per $1 TCO vs. Interactivity