GB300 NVL72 FP4: MTP vs Off Speculative Decoding
Speculative decoding comparison of MTP versus Off on GB300 NVL72 FP4 (NVIDIA Blackwell) running GLM 5/5.1. 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 21–104 tok/s/user interactivity band, at 42 tok/s/user on GLM 5/5.1 (GB300 NVL72 FP4): MTP runs 10234 tok/s/GPU at $0.07/M tokens, Off runs 5653 at $0.14/M. MTP is 91% cheaper per token; MTP delivers 81% more tok/s/GPU. Gains from speculative decoding vary by workload; short-output prompts tend to benefit less.
At 63 tok/s/user on GLM 5/5.1 (GB300 NVL72 FP4), MTP delivers 7527 tok/s/GPU at $0.10 per million tokens; Off delivers 429 tok/s/GPU at $1.71. MTP is 1666% cheaper per token; MTP delivers 1657% more tok/s/GPU. Speculative decoding accepts draft tokens to reduce per-token latency — gains vary by workload and prompt distribution.
MTP posts 4517 tok/s/GPU for $0.16 per million tokens at 84 tok/s/user on GLM 5/5.1 (GB300 NVL72 FP4); Off posts 92 tok/s/GPU for $7.95. MTP is 4869% cheaper per token; MTP delivers 4832% more tok/s/GPU. Draft-token acceptance rates determine whether speculative decoding helps or hurts at a given concurrency level. (Numbers reflect this URL's pinned 1k/1k · fp4 workload — changing sequence or model updates both the table and chart; the table stays pinned to this page's precision, so precision toggles in the controls affect the chart only.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/gpu) | MTP:10233.9Off:5653.3 | MTP:7526.9Off:428.5 | MTP:4516.5Off:91.6 |
| Cost ($/M tok) | MTP:$0.072Off:$0.138 | MTP:$0.097Off:$1.713 | MTP:$0.160Off:$7.954 |
| tok/s/MW | MTP:4873292Off:2692065 | MTP:3584260Off:204052 | MTP:2150736Off:43610 |
| Concurrency | MTP:~3770Off:~2253 | MTP:~1486Off:~104 | MTP:~721Off:~35 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.