Head-to-head AI inference benchmark comparison of B200 (NVIDIA Blackwell) and B300 (NVIDIA Blackwell) on DeepSeek V4.1 Flash 552B. 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 →
Setting 106 tok/s/user as the target on DeepSeek V4.1 Flash 552B, B200 produces 58063 tok/s/chip ($0.01 per million tokens) and B300 produces 55501 ($0.01). B200 is 37% cheaper per token; B200 delivers 5% more tok/s/chip.
At 186 tok/s/user interactivity on DeepSeek V4.1 Flash 552B, B200 delivers 26374 tok/s/chip at $0.02 per million tokens; B300 delivers 25068 tok/s/chip at $0.03. B200 is 37% cheaper per token; B200 delivers 5% more tok/s/chip at this point.
B200 posts 14033 tok/s/chip for $0.03 per million tokens at 267 tok/s/user on DeepSeek V4.1 Flash 552B; B300 posts 13705 tok/s/chip for $0.05. B200 is 34% cheaper per token; B200 delivers 2% more tok/s/chip. (Numbers reflect the default agentic-traces · fp4 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | B200:58062.9B300:55501.1 | B200:26374.3B300:25067.9 | B200:14033.1B300:13704.9 |
| Cost ($/M tok) | B200:$0.008B300:$0.011 | B200:$0.018B300:$0.025 | B200:$0.034B300:$0.046 |
| tok/s/MW | B200:33954924B300:29211101 | B200:15423596B300:13193641 | B200:8206500B300:7213090 |
| Concurrency | B200:~33B300:~31 | B200:~15B300:~14 | B200:~7B300:~7 |
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.
No measurements to plot for this selection. Review the benchmark controls above or adjust quick filters.
