DeepSeek V4 Pro 0813 1.6T · Chip comparison

DeepSeek V4 Pro 0813 1.6T — B300 vs GB200 NVL72

Head-to-head AI inference benchmark comparison of B300 (NVIDIA Blackwell) and GB200 NVL72 (NVIDIA Blackwell) on DeepSeek V4 Pro 0813 1.6T. 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 113 tok/s/user interactivity on DeepSeek V4 Pro 0813 1.6T, B300 delivers 27,740 tok/s/chip at $0.023 per million tokens; GB200 NVL72 delivers 30,877 tok/s/chip at $0.017. GB200 NVL72 is 26% cheaper per token; GB200 NVL72 delivers 11% more tok/s/chip at this point.

B300 posts 10,191 tok/s/chip for $0.062 per million tokens at 171 tok/s/user on DeepSeek V4 Pro 0813 1.6T; GB200 NVL72 posts 12,467 tok/s/chip for $0.041. GB200 NVL72 is 33% cheaper per token; GB200 NVL72 delivers 22% more tok/s/chip.

Throughput at 229 tok/s/user on DeepSeek V4 Pro 0813 1.6T: B300 hits 3,512 tok/s/chip, GB200 NVL72 hits 4,622. Costs per million tokens land at $0.18 and $0.11 respectively. GB200 NVL72 is 37% cheaper per token; GB200 NVL72 delivers 32% 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.)

View performance per dollar →

Interpolated from real benchmark data. Edit target interactivity values below to compare at different operating points.

Metric
Interactivity (tok/s/user)
Interactivity (tok/s/user)
Interactivity (tok/s/user)
Throughput (tok/s/chip)
B30027,740GB200 NVL7230,877
B30010,191GB200 NVL7212,467
B3003,512GB200 NVL724,622
Cost ($/M tok)
B300$0.023GB200 NVL72$0.017
B300$0.062GB200 NVL72$0.041
B300$0.18GB200 NVL72$0.11
tok/s/MW
B30014,600,163GB200 NVL7216,511,700
B3005,363,669GB200 NVL726,666,914
B3001,848,455GB200 NVL722,471,559
Concurrency
B300~69GB200 NVL72~198
B300~5GB200 NVL72~41
B300~4GB200 NVL72~6