Qwen 3.5 397B-A17B · MI355X · Precision comparison

MI355X: FP4 vs BF16 precision comparison

How FP4 and BF16 precision affect Qwen 3.5 397B-A17B inference on MI355X (AMD CDNA 4). Throughput, latency, and cost across LLM workloads. Use the chart controls below to switch sequences and metrics — same interactions as the main inference chart.

At 64 tok/s/user on Qwen 3.5 397B-A17B (MI355X), FP4 delivers 2,220 tok/s/chip at $0.19 per million tokens; BF16 delivers 879 tok/s/chip at $0.47. FP4 delivers 153% more tok/s/chip on the same chip, which makes it 60% cheaper per token. Lower-precision quantization trades model accuracy for throughput — check the evaluation page for quality impact.

FP4 posts 1,475 tok/s/chip for $0.28 per million tokens at 99 tok/s/user on Qwen 3.5 397B-A17B (MI355X); BF16 posts 597 tok/s/chip for $0.70. FP4 delivers 147% more tok/s/chip on the same chip, which makes it 60% cheaper per token. Quantization-level accuracy differences are tracked on the evaluation tab.

Throughput at 133 tok/s/user on Qwen 3.5 397B-A17B (MI355X): FP4 hits 1,104 tok/s/chip, BF16 hits 279. Costs per million tokens land at $0.38 and $1.49 respectively. FP4 delivers 4.0× the tok/s/chip on the same chip, which makes it 75% cheaper per token. The cost-throughput tradeoff from lower precision is only part of the picture — see the evaluation page for accuracy data. (Numbers reflect the default 1k/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)
FP42,220BF16879
FP41,475BF16597
FP41,104BF16279
Cost ($/M tok)
FP4$0.19BF16$0.47
FP4$0.28BF16$0.70
FP4$0.38BF16$1.49
tok/s/MW
FP41,062,420BF16420,444
FP4705,750BF16285,490
FP4528,455BF16133,623
Concurrency
FP4~36BF16~128
FP4~15BF16~24
FP4~9BF16~9
Qwen 3.5 397B-A17B: FP4 versus BF16 throughput and cost at matched interactivity levels on MI355X
FP4 versus BF16 throughput and cost per million tokens on MI355X for this comparison's canonical default workload
Open full-size image ↗