DeepSeek R1 · GB200 NVL72 · Precision comparison

GB200 NVL72: FP4 vs FP8 precision comparison

How FP4 and FP8 precision affect DeepSeek R1 inference on GB200 NVL72 (NVIDIA Blackwell). Throughput, latency, and cost across LLM workloads. Use the chart controls below to switch sequences and metrics — same interactions as the main inference chart.

FP4 posts 12,458 tok/s/chip for $0.041 per million tokens at 76 tok/s/user on DeepSeek R1 (GB200 NVL72); FP8 posts 5,749 tok/s/chip for $0.090. FP4 delivers 117% more tok/s/chip on the same chip, which makes it 54% cheaper per token. Quantization-level accuracy differences are tracked on the evaluation tab.

Throughput at 120 tok/s/user on DeepSeek R1 (GB200 NVL72): FP4 hits 9,356 tok/s/chip, FP8 hits 3,988. Costs per million tokens land at $0.055 and $0.13 respectively. FP4 delivers 135% more tok/s/chip on the same chip, which makes it 57% cheaper per token. The cost-throughput tradeoff from lower precision is only part of the picture — see the evaluation page for accuracy data.

Toward the upper edge of the 32–208 tok/s/user interactivity band, at 164 tok/s/user on DeepSeek R1 (GB200 NVL72): FP4 runs 3,995 tok/s/chip at $0.13 per million tokens, FP8 runs 818 at $0.63. FP4 delivers 4.9× the tok/s/chip on the same chip, which makes it 80% cheaper per token. Precision changes affect both inference speed and model quality — consult the evaluation tab for accuracy benchmarks. (Numbers reflect the default 8k/1k selection for this URL — table and chart below update if you change sequence or model in the controls. Each side uses the best available serving configuration for that precision, which may include speculative decoding such as MTP where recipes exist — the same convention as the other comparison pages.)

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)
FP412,458FP85,749
FP49,356FP83,988
FP43,995FP8818
Cost ($/M tok)
FP4$0.041FP8$0.090
FP4$0.055FP8$0.13
FP4$0.13FP8$0.63
tok/s/MW
FP46,662,190FP83,074,446
FP45,003,271FP82,132,361
FP42,136,436FP8437,312
Concurrency
FP4~1,076FP8~618
FP4~721FP8~333
FP4~164FP8~27
DeepSeek R1: FP4 versus FP8 throughput and cost at matched interactivity levels on GB200 NVL72
FP4 versus FP8 throughput and cost per million tokens on GB200 NVL72 for this comparison's canonical default workload
Open full-size image ↗