Kimi K2.5/K2.6/K2.7-Code 1T — B300 vs GB200 NVL72
Head-to-head AI inference benchmark comparison of B300 (NVIDIA Blackwell) and GB200 NVL72 (NVIDIA Blackwell) on Kimi K2.5/K2.6/K2.7-Code 1T. 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 62 tok/s/user interactivity on Kimi K2.5/K2.6/K2.7-Code 1T, B300 delivers 2929 tok/s/chip at $0.21 per million tokens; GB200 NVL72 delivers 9455 tok/s/chip at $0.05. GB200 NVL72 is 292% cheaper per token; GB200 NVL72 delivers 223% more tok/s/chip at this point.
B300 posts 1731 tok/s/chip for $0.36 per million tokens at 99 tok/s/user on Kimi K2.5/K2.6/K2.7-Code 1T; GB200 NVL72 posts 1548 tok/s/chip for $0.33. GB200 NVL72 is 9% cheaper per token; B300 delivers 12% more tok/s/chip.
Throughput at 136 tok/s/user on Kimi K2.5/K2.6/K2.7-Code 1T: B300 hits 783 tok/s/chip, GB200 NVL72 hits 837. Per-million costs land at $0.80 and $0.62 respectively. GB200 NVL72 is 30% cheaper per token; GB200 NVL72 delivers 7% more tok/s/chip. (Numbers reflect the default 8k/1k · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B300:2929.1GB200 NVL72:9454.9 | B300:1731.4GB200 NVL72:1548.2 | B300:783.1GB200 NVL72:836.8 |
| Cost ($/M tok) | B300:$0.214GB200 NVL72:$0.055 | B300:$0.363GB200 NVL72:$0.334 | B300:$0.802GB200 NVL72:$0.617 |
| tok/s/MW | B300:1541639GB200 NVL72:5056105 | B300:911253GB200 NVL72:827888 | B300:412168GB200 NVL72:447506 |
| Concurrency | B300:~21GB200 NVL72:~1229 | B300:~8GB200 NVL72:~256 | B300:~3GB200 NVL72:~28 |
Inference Performance
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.
Total Tokens per $1 USD (Owning - Hyperscaler) vs. Interactivity
Kimi K2.5/2.6/2.7-Code 1T • FP4 • 8K / 1K • Source: SemiAnalysis InferenceX™
TCO $/chip/hr: H100: 1.17H200: 1.22B200: 1.73B300: 2.26GB200: 1.86GB300: 2.31MI300X: 0.95MI325X: 1.1MI355X: 1.5RTX6000PRO: 0.68
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