MI355X: FP4 vs FP8 Precision Comparison

FP4 posts 2950 tok/s/GPU for $0.14 per million tokens at 31 tok/s/user on DeepSeek V4 Pro 1.6T (MI355X); FP8 posts 20 tok/s/GPU for $20.1. FP4 is 14306% cheaper per token; FP4 delivers 14356% more tok/s/GPU. Quantization-level accuracy differences are tracked on the evaluation tab.

Throughput at 60 tok/s/user on DeepSeek V4 Pro 1.6T (MI355X): FP4 hits 1106 tok/s/GPU, FP8 hits 20. Per-million costs land at $0.37 and $20.1 respectively. FP4 is 5351% cheaper per token; FP4 delivers 5322% more tok/s/GPU. 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 1–120 tok/s/user interactivity band, at 90 tok/s/user on DeepSeek V4 Pro 1.6T (MI355X): FP4 runs 616 tok/s/GPU at $0.68/M tokens, FP8 runs 20 at $20.1/M. FP4 is 2867% cheaper per token; FP4 delivers 2919% more tok/s/GPU. 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.)