H100 FP8: MTP vs Off Speculative Decoding
Speculative decoding comparison of MTP versus Off on H100 FP8 (NVIDIA Hopper) 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.
Near the low end of the 60–157 tok/s/user interactivity band, at 84 tok/s/user on Qwen 3.5 397B-A17B (H100 FP8): MTP runs 361 tok/s/chip at $0.90/M tokens, Off runs 214 at $1.52/M. MTP is 69% cheaper per token; MTP delivers 69% more tok/s/chip. Gains from speculative decoding vary by workload; short-output prompts tend to benefit less.
At 108 tok/s/user on Qwen 3.5 397B-A17B (H100 FP8), MTP delivers 292 tok/s/chip at $1.11 per million tokens; Off delivers 160 tok/s/chip at $2.03. MTP is 82% cheaper per token; MTP delivers 82% more tok/s/chip. Speculative decoding accepts draft tokens to reduce per-token latency — gains vary by workload and prompt distribution.
MTP posts 240 tok/s/chip for $1.35 per million tokens at 133 tok/s/user on Qwen 3.5 397B-A17B (H100 FP8); Off posts 102 tok/s/chip for $3.20. MTP is 137% cheaper per token; MTP delivers 137% more tok/s/chip. Draft-token acceptance rates determine whether speculative decoding helps or hurts at a given concurrency level. (Numbers reflect this URL's pinned 1k/1k · fp8 workload — changing sequence or model updates both the table and chart; the table stays pinned to this page's precision, so precision toggles in the controls affect the chart only.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/chip) | MTP:360.9Off:213.9 | MTP:291.9Off:160.2 | MTP:240.5Off:101.5 |
| Cost ($/M tok) | MTP:$0.901Off:$1.519 | MTP:$1.113Off:$2.029 | MTP:$1.352Off:$3.201 |
| tok/s/MW | MTP:263401Off:156145 | MTP:213065Off:116900 | MTP:175528Off:74101 |
| Concurrency | MTP:~18Off:~11 | MTP:~11Off:~6 | MTP:~8Off:~3 |
Inference Performance
Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.