GB200 NVL72 FP8: MTP vs Off Speculative Decoding
Speculative decoding comparison of MTP versus Off on GB200 NVL72 FP8 (NVIDIA Blackwell) running Qwen 3.5 397B-A17B. Throughput, cost, and interactivity differences across LLM workloads. Use the chart controls below to switch sequences and metrics — same interactions as the main inference chart.
MTP acceptance-rate implementations differ across inference engines. Points from different engines are not directly comparable on the same curve — throughput and cost at matched interactivity may reflect engine-level differences rather than pure speculative decoding gains. Interpret cross-engine comparisons with caution.
At 89 tok/s/user on Qwen 3.5 397B-A17B (GB200 NVL72 FP8), MTP delivers 10304 tok/s/chip at $0.05 per million tokens; Off delivers 3468 tok/s/chip at $0.15. MTP is 197% cheaper per token; MTP delivers 197% more tok/s/chip. Speculative decoding accepts draft tokens to reduce per-token latency — gains vary by workload and prompt distribution.
MTP posts 7708 tok/s/chip for $0.07 per million tokens at 130 tok/s/user on Qwen 3.5 397B-A17B (GB200 NVL72 FP8); Off posts 1855 tok/s/chip for $0.28. MTP is 316% cheaper per token; MTP delivers 316% more tok/s/chip. Draft-token acceptance rates determine whether speculative decoding helps or hurts at a given concurrency level.
Throughput at 172 tok/s/user on Qwen 3.5 397B-A17B (GB200 NVL72 FP8): MTP hits 3808 tok/s/chip, Off hits 939. Per-million costs land at $0.14 and $0.55 respectively. MTP is 306% cheaper per token; MTP delivers 306% more tok/s/chip. Speculative decoding trades extra compute on draft tokens for fewer decoding steps — the payoff depends on sequence length and batch size. (Numbers reflect this URL's pinned 8k/1k · fp8 workload — changing sequence or model updates both the table and chart; the table stays pinned to this page's precision, so precision toggles in the controls affect the chart only.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | MTP:10304.0Off:3467.7 | MTP:7708.5Off:1854.9 | MTP:3808.3Off:938.7 |
| Cost ($/M tok) | MTP:$0.050Off:$0.149 | MTP:$0.067Off:$0.279 | MTP:$0.136Off:$0.550 |
| tok/s/MW | MTP:5510151Off:1854362 | MTP:4122183Off:991918 | MTP:2036551Off:502002 |
| Concurrency | MTP:~1083Off:~38 | MTP:~437Off:~13 | MTP:~50Off:~5 |
Inference Performance
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.