DeepSeek R1 · B200 · Precision comparison
B200: FP4 vs FP8 precision comparison
How FP4 and FP8 precision affect DeepSeek R1 inference on B200 (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 9,367 tok/s/chip for $0.051 per million tokens at 77 tok/s/user on DeepSeek R1 (B200); FP8 posts 2,297 tok/s/chip for $0.21. FP4 delivers 4.1× the tok/s/chip on the same chip, which makes it 75% cheaper per token. Quantization-level accuracy differences are tracked on the evaluation tab.
Throughput at 138 tok/s/user on DeepSeek R1 (B200): FP4 hits 2,267 tok/s/chip, FP8 hits 989. Costs per million tokens land at $0.21 and $0.49 respectively. FP4 delivers 129% more tok/s/chip on the same chip, which makes it 56% 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 16–260 tok/s/user interactivity band, at 200 tok/s/user on DeepSeek R1 (B200): FP4 runs 1,400 tok/s/chip at $0.34 per million tokens, FP8 runs 352 at $1.37. FP4 delivers 4.0× the tok/s/chip on the same chip, which makes it 75% cheaper per token. Precision changes affect both inference speed and model quality — consult the evaluation tab for accuracy benchmarks. (Numbers reflect the default 8k/1k selection for this URL — table and chart below update if you change sequence or model in the controls. Each side uses the best available serving configuration for that precision, which may include speculative decoding such as MTP where recipes exist — the same convention as the other comparison pages.)
Interpolated from real benchmark data. Edit target interactivity values below to compare at different operating points.
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | FP49,367FP82,297 | FP42,267FP8989 | FP41,400FP8352 |
| Cost ($/M tok) | FP4$0.051FP8$0.21 | FP4$0.21FP8$0.49 | FP4$0.34FP8$1.37 |
| tok/s/MW | FP45,477,576FP81,343,043 | FP41,325,560FP8578,501 | FP4818,993FP8205,871 |
| Concurrency | FP4~655FP8~123 | FP4~12FP8~19 | FP4~60FP8~2 |
