Kimi K3 2.8T — H200 vs TPU7x Performance per Dollar
Cost per million tokens of H200 (NVIDIA Hopper) versus TPU7x (Google Ironwood) on Kimi K3 2.8T. Large-hyperscaler-volume ownership TCO normalized by total 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.
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 →
Push Kimi K3 2.8T to 99 tok/s/user and H200 lands at $4.85 per million tokens against TPU7x's $0.06 — TPU7x pulls ahead by 7870%.
H200: $4.85 per million tokens. TPU7x: $0.18. Both at 196 tok/s/user on Kimi K3 2.8T, with TPU7x 2626% cheaper.
Toward the upper edge of the 2–390 tok/s/user interactivity band — at 293 tok/s/user — H200 runs $4.85 per million tokens on Kimi K3 2.8T while TPU7x runs $0.30. TPU7x is the cheaper choice by 1539%. (Numbers reflect the default agentic-traces · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
Chip pricing (owning hyperscaler): H200 $1.22/chip/hr · TPU7x $1.03/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 | H200:$4.850TPU7x:$0.061 | H200:$4.850TPU7x:$0.178 | H200:$4.850TPU7x:$0.296 |
| Concurrency | H200:~1TPU7x:~26 | H200:~1TPU7x:~7 | H200:~1TPU7x:~3 |
Inference Performance
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.
No data available
No measurements to plot for this selection. Review the benchmark controls above or adjust quick filters.