Qwen 3.5 397B-A17B · MI355X · Precision comparison
MI355X: FP4 vs FP8 precision comparison
How FP4 and FP8 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.
FP4 posts 2,158 tok/s/chip for $0.19 per million tokens at 68 tok/s/user on Qwen 3.5 397B-A17B (MI355X); FP8 posts 2,012 tok/s/chip for $0.21. FP4 delivers 7% more tok/s/chip on the same chip, which makes it 7% cheaper per token. Quantization-level accuracy differences are tracked on the evaluation tab.
Throughput at 119 tok/s/user on Qwen 3.5 397B-A17B (MI355X): FP4 hits 1,229 tok/s/chip, FP8 hits 1,050. Costs per million tokens land at $0.34 and $0.40 respectively. FP4 delivers 17% more tok/s/chip on the same chip, which makes it 15% 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 17–220 tok/s/user interactivity band, at 170 tok/s/user on Qwen 3.5 397B-A17B (MI355X): FP4 runs 757 tok/s/chip at $0.55 per million tokens, FP8 runs 590 at $0.71. FP4 delivers 28% more tok/s/chip on the same chip, which makes it 22% cheaper per token. Precision changes affect both inference speed and model quality — consult the evaluation tab for accuracy benchmarks. (Numbers reflect the default 1k/1k selection for this URL — table and chart below update if you change sequence or model in the controls. Each side uses the best available serving configuration for that precision, which may include speculative decoding such as MTP where recipes exist — the same convention as the other comparison pages.)
Interpolated from real benchmark data. Edit target interactivity values below to compare at different operating points.
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | FP42,158FP82,012 | FP41,229FP81,050 | FP4757FP8590 |
| Cost ($/M tok) | FP4$0.19FP8$0.21 | FP4$0.34FP8$0.40 | FP4$0.55FP8$0.71 |
| tok/s/MW | FP41,032,553FP8962,539 | FP4587,966FP8502,164 | FP4362,046FP8282,332 |
| Concurrency | FP4~33FP8~59 | FP4~11FP8~18 | FP4~4FP8~8 |
