DeepSeek R1 · Performance per Dollar

DeepSeek R1 — B200 vs H200 Performance per Dollar

Cost per million tokens of B200 (NVIDIA Blackwell) versus H200 (NVIDIA Hopper) on DeepSeek R1. 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.

B200 edges H200 at 59 tok/s/user on DeepSeek R1 — $0.20 per million tokens versus $0.93, a 365% cost-per-token gap.

Push DeepSeek R1 to 98 tok/s/user and B200 lands at $1.04 per million tokens against H200's $2.66 — B200 pulls ahead by 157%.

B200: $2.13 per million tokens. H200: $9.94. Both at 136 tok/s/user on DeepSeek R1, with B200 367% cheaper. (Numbers reflect the default 1k/1k · fp8 selection for this URL — table and chart below update if you change sequence, precision, or model in the controls.)

GPU pricing (owning hyperscaler): B200 $1.95/GPU/hr · H200 $1.41/GPU/hr. Source: SemiAnalysis Market August 2025 Pricing Surveys & AI Cloud TCO Model.

View full latency + throughput comparison →

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)
Dollar per Million Tokens
B200:$0.200H200:$0.929
B200:$1.037H200:$2.663
B200:$2.127H200:$9.940
Concurrency
B200:~1030H200:~133
B200:~185H200:~6
B200:~13H200:~14

Inference Performance

Inference performance metrics across different models, hardware configurations, and serving parameters.