Kimi K3 2.8T — B300 vs GB200 NVL72 Performance per Dollar
Cost per million tokens of B300 (NVIDIA Blackwell) versus GB200 NVL72 (NVIDIA Blackwell) on Kimi K3 2.8T. 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.
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 →
On Kimi K3 2.8T at 55 tok/s/user, the per-million math comes out to $0.11 for B300 and $0.12 for GB200 NVL72; B300 delivers 12% more output per dollar.
At 103 tok/s/user on Kimi K3 2.8T, B300 costs $0.16 per million tokens; GB200 NVL72 costs $0.24. B300 is 50% more cost-efficient at this operating point.
B300 edges GB200 NVL72 at 150 tok/s/user on Kimi K3 2.8T — $0.30 per million tokens versus $0.41, a 37% cost-per-token gap. (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): B300 $2.26/chip/hr · GB200 NVL72 $1.86/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 | B300:$0.109GB200 NVL72:$0.123 | B300:$0.158GB200 NVL72:$0.238 | B300:$0.303GB200 NVL72:$0.414 |
| Concurrency | B300:~9GB200 NVL72:~16 | B300:~6GB200 NVL72:~8 | B300:~2GB200 NVL72:~3 |
Inference Performance
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.