DeepSeek V4.1 Flash 552B · Chip comparison

DeepSeek V4.1 Flash 552B — GB300 NVL72 vs Vera Rubin NVL72

Head-to-head AI inference benchmark comparison of GB300 NVL72 (NVIDIA Blackwell) and Vera Rubin NVL72 (NVIDIA Vera Rubin) 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 →

Throughput at 567 tok/s/user on DeepSeek V4.1 Flash 552B: GB300 NVL72 hits 52,284 tok/s/chip, Vera Rubin NVL72 hits 1,078,245. Costs per million tokens land at $0.012 and $0.001 respectively. Vera Rubin NVL72 is 92% cheaper per token; Vera Rubin NVL72 delivers 21× the tok/s/chip.

GB300 NVL72 / Vera Rubin NVL72 on DeepSeek V4.1 Flash 552B at 952 tok/s/user: 8,108 / 128,998 tok/s/chip, $0.079 / $0.008 per million tokens. Vera Rubin NVL72 is 90% cheaper per token; Vera Rubin NVL72 delivers 16× the tok/s/chip.

Toward the upper edge of the 181–1724 tok/s/user interactivity band, at 1338 tok/s/user on DeepSeek V4.1 Flash 552B: GB300 NVL72 runs 4,448 tok/s/chip at $0.14 per million tokens, Vera Rubin NVL72 runs 51,010 at $0.020. Vera Rubin NVL72 is 86% cheaper per token; Vera Rubin NVL72 delivers 11× the 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)
GB300 NVL7252,284Vera Rubin NVL721,078,245
GB300 NVL728,108Vera Rubin NVL72128,998
GB300 NVL724,448Vera Rubin NVL7251,010
Cost ($/M tok)
GB300 NVL72$0.012Vera Rubin NVL72$0.001
GB300 NVL72$0.079Vera Rubin NVL72$0.008
GB300 NVL72$0.14Vera Rubin NVL72$0.020
tok/s/MW
GB300 NVL7224,662,049Vera Rubin NVL72326,741,042
GB300 NVL723,824,625Vera Rubin NVL7239,090,309
GB300 NVL722,098,063Vera Rubin NVL7215,457,720
Concurrency
GB300 NVL72~25Vera Rubin NVL72~6,459
GB300 NVL72~1Vera Rubin NVL72~276
GB300 NVL72~1Vera Rubin NVL72~157