Kimi K3.1 Pro — B200 vs GB300 NVL72
Head-to-head AI inference benchmark comparison of B200 (NVIDIA Blackwell) and GB300 NVL72 (NVIDIA Blackwell) on Kimi K3.1 Pro. 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.
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 →
At 67 tok/s/user interactivity on Kimi K3.1 Pro, B200 delivers 1590 tok/s/chip at $0.30 per million tokens; GB300 NVL72 delivers 29513 tok/s/chip at $0.02. GB300 NVL72 is 1290% cheaper per token; GB300 NVL72 delivers 1756% more tok/s/chip at this point.
B200 posts 1450 tok/s/chip for $0.33 per million tokens at 72 tok/s/user on Kimi K3.1 Pro; GB300 NVL72 posts 27813 tok/s/chip for $0.02. GB300 NVL72 is 1337% cheaper per token; GB300 NVL72 delivers 1818% more tok/s/chip.
Throughput at 78 tok/s/user on Kimi K3.1 Pro: B200 hits 1122 tok/s/chip, GB300 NVL72 hits 25614. Per-million costs land at $0.43 and $0.03 respectively. GB300 NVL72 is 1609% cheaper per token; GB300 NVL72 delivers 2182% more tok/s/chip. (Numbers reflect the default agentic-traces · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B200:1590.1GB300 NVL72:29512.8 | B200:1449.7GB300 NVL72:27812.6 | B200:1122.2GB300 NVL72:25613.6 |
| Cost ($/M tok) | B200:$0.302GB300 NVL72:$0.022 | B200:$0.331GB300 NVL72:$0.023 | B200:$0.428GB300 NVL72:$0.025 |
| tok/s/MW | B200:929863GB300 NVL72:13921136 | B200:847792GB300 NVL72:13119142 | B200:656246GB300 NVL72:12081897 |
| Concurrency | B200:~6GB300 NVL72:~256 | B200:~5GB300 NVL72:~218 | B200:~3GB300 NVL72:~176 |