Qwen 3.5 397B-A17B · MI325X · Precision comparison

MI325X: FP8 vs BF16 precision comparison

How FP8 and BF16 precision affect Qwen 3.5 397B-A17B inference on MI325X (AMD CDNA 3). Throughput, latency, and cost across LLM workloads. Use the chart controls below to switch sequences and metrics — same interactions as the main inference chart.

Near the low end of the 41–72 tok/s/user interactivity band, at 49 tok/s/user on Qwen 3.5 397B-A17B (MI325X): FP8 runs 347 tok/s/chip at $0.88 per million tokens, BF16 runs 462 at $0.66. BF16 delivers 33% more tok/s/chip on the same chip, which makes it 25% cheaper per token. Precision changes affect both inference speed and model quality — consult the evaluation tab for accuracy benchmarks.

At 57 tok/s/user on Qwen 3.5 397B-A17B (MI325X), FP8 delivers 211 tok/s/chip at $1.45 per million tokens; BF16 delivers 382 tok/s/chip at $0.80. BF16 delivers 81% more tok/s/chip on the same chip, which makes it 45% cheaper per token. Lower-precision quantization trades model accuracy for throughput — check the evaluation page for quality impact.

FP8 posts 124 tok/s/chip for $2.46 per million tokens at 65 tok/s/user on Qwen 3.5 397B-A17B (MI325X); BF16 posts 307 tok/s/chip for $1.00. BF16 delivers 147% more tok/s/chip on the same chip, which makes it 59% cheaper per token. Quantization-level accuracy differences are tracked on the evaluation tab. (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)
FP8347BF16462
FP8211BF16382
FP8124BF16307
Cost ($/M tok)
FP8$0.88BF16$0.66
FP8$1.45BF16$0.80
FP8$2.46BF16$1.00
tok/s/MW
FP8205,060BF16273,085
FP8124,837BF16226,025
FP873,546BF16181,594
Concurrency
FP8~30BF16~41
FP8~16BF16~30
FP8~8BF16~20
Qwen 3.5 397B-A17B: FP8 versus BF16 throughput and cost at matched interactivity levels on MI325X
FP8 versus BF16 throughput and cost per million tokens on MI325X for this comparison's canonical default workload
Open full-size image ↗