Kimi K2.5/K2.6/K2.7-Code 1T — B300 vs H200
Head-to-head AI inference benchmark comparison of B300 (NVIDIA Blackwell) and H200 (NVIDIA Hopper) 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 40 tok/s/user interactivity on Kimi K2.5/K2.6/K2.7-Code 1T, B300 delivers 2181 tok/s/chip at $0.29 per million tokens; H200 delivers 867 tok/s/chip at $0.39. B300 is 36% cheaper per token; B300 delivers 151% more tok/s/chip at this point.
B300 posts 1575 tok/s/chip for $0.40 per million tokens at 58 tok/s/user on Kimi K2.5/K2.6/K2.7-Code 1T; H200 posts 700 tok/s/chip for $0.48. B300 is 22% cheaper per token; B300 delivers 125% more tok/s/chip.
Throughput at 75 tok/s/user on Kimi K2.5/K2.6/K2.7-Code 1T: B300 hits 1173 tok/s/chip, H200 hits 548. Per-million costs land at $0.54 and $0.62 respectively. B300 is 16% cheaper per token; B300 delivers 114% more tok/s/chip. (Numbers reflect the default 8k/1k · int4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B300:2180.9H200:867.4 | B300:1575.0H200:699.6 | B300:1173.2H200:547.6 |
| Cost ($/M tok) | B300:$0.288H200:$0.391 | B300:$0.399H200:$0.484 | B300:$0.535H200:$0.619 |
| tok/s/MW | B300:1147838H200:633138 | B300:828927H200:510677 | B300:617461H200:399674 |
| Concurrency | B300:~24H200:~20 | B300:~12H200:~11 | B300:~7H200:~7 |
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