Head-to-head AI inference benchmark comparison of GB200 NVL72 (NVIDIA Blackwell) and GB300 NVL72 (NVIDIA Blackwell) on MiniMax M2.5/M2.7. 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.
At 48 tok/s/user interactivity on MiniMax M2.5/M2.7, GB200 NVL72 delivers 8970 tok/s/chip at $0.06 per million tokens; GB300 NVL72 delivers 9393 tok/s/chip at $0.07. GB200 NVL72 is 19% cheaper per token; GB300 NVL72 delivers 5% more tok/s/chip at this point.
GB200 NVL72 posts 5557 tok/s/chip for $0.09 per million tokens at 65 tok/s/user on MiniMax M2.5/M2.7; GB300 NVL72 posts 5573 tok/s/chip for $0.12. GB200 NVL72 is 24% cheaper per token; throughput per chip is essentially tied.
Throughput at 83 tok/s/user on MiniMax M2.5/M2.7: GB200 NVL72 hits 3120 tok/s/chip, GB300 NVL72 hits 3547. Per-million costs land at $0.17 and $0.18 respectively. GB200 NVL72 is 9% cheaper per token; GB300 NVL72 delivers 14% more tok/s/chip. (Numbers reflect the default 8k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | GB200 NVL72:8969.9GB300 NVL72:9393.1 | GB200 NVL72:5556.6GB300 NVL72:5573.4 | GB200 NVL72:3120.4GB300 NVL72:3546.8 |
| Cost ($/M tok) | GB200 NVL72:$0.058GB300 NVL72:$0.068 | GB200 NVL72:$0.093GB300 NVL72:$0.115 | GB200 NVL72:$0.166GB300 NVL72:$0.181 |
| tok/s/MW | GB200 NVL72:4796740GB300 NVL72:4430716 | GB200 NVL72:2971457GB300 NVL72:2628982 | GB200 NVL72:1668687GB300 NVL72:1673039 |
| Concurrency | GB200 NVL72:~146GB300 NVL72:~136 | GB200 NVL72:~64GB300 NVL72:~64 | GB200 NVL72:~27GB300 NVL72:~48 |
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.
TCO $/chip/hr: VR200: 3.61H100: 1.17H200: 1.22B200: 1.73B300: 2.26GB200: 1.86GB300: 2.31MI300X: 0.95MI325X: 1.1MI355X: 1.5RTX6000PRO: 0.68Jalapeño (Teacup): 1.47
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 • Drag to pan • Double-click to reset • Click a point to pin tooltip
