Kimi K2.5/K2.6/K2.7-Code 1T — B200 vs H200 Performance per Dollar
Cost per million tokens of B200 (NVIDIA Blackwell) versus H200 (NVIDIA Hopper) on Kimi K2.5/K2.6/K2.7-Code 1T. Owning-hyperscaler TCO normalized by output tokens — performance per dollar across LLM workloads. Pick the more cost-efficient SKU at every target interactivity level. Use the chart controls below to switch sequences, precisions, and metrics — same interactions as the main inference chart.
Push Kimi K2.5/K2.6/K2.7-Code 1T to 54 tok/s/user and B200 lands at $1.36 per million tokens against H200's $1.13 — H200 pulls ahead by 20%.
B200: $2.02 per million tokens. H200: $1.53. Both at 71 tok/s/user on Kimi K2.5/K2.6/K2.7-Code 1T, with H200 32% cheaper.
Toward the upper edge of the 37–105 tok/s/user interactivity band — at 88 tok/s/user — B200 runs $2.86 per million tokens on Kimi K2.5/K2.6/K2.7-Code 1T while H200 runs $2.16. H200 is the cheaper choice by 32%. (Numbers reflect the default 1k/1k · int4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
Chip pricing (owning hyperscaler): B200 $1.73/chip/hr · H200 $1.22/chip/hr. Source: SemiAnalysis Market July 2026 Pricing Surveys & AI Cloud TCO Model.
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Dollar per Million Tokens | B200:$1.358H200:$1.130 | B200:$2.018H200:$1.526 | B200:$2.856H200:$2.157 |
| Concurrency | B200:~26H200:~22 | B200:~13H200:~12 | B200:~8H200:~7 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.