B300 FP4: MTP vs Off Speculative Decoding
Speculative decoding comparison of MTP versus Off on B300 FP4 (NVIDIA Blackwell) running Kimi K2.5/K2.6/K2.7-Code 1T. 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.
Off hits 932 tok/s/GPU for $0.70 per million tokens at 69 tok/s/user on Kimi K2.5/K2.6/K2.7-Code 1T (B300 FP4). No MTP data at this operating point.
Off: 899 tok/s/GPU, $0.72 per million tokens at 104 tok/s/user on Kimi K2.5/K2.6/K2.7-Code 1T (B300 FP4). MTP is unmeasured here.
At 138 tok/s/user on Kimi K2.5/K2.6/K2.7-Code 1T (B300 FP4), Off delivers 493 tok/s/GPU at $1.32 per million tokens; MTP hasn't been benchmarked at this target. (Numbers reflect this URL's pinned 1k/1k · fp4 workload — changing sequence or model updates both the table and chart; the table stays pinned to this page's precision, so precision toggles in the controls affect the chart only.)
| Metric | Interactivity (tok/s/user) | Interactivity (tok/s/user) | Interactivity (tok/s/user) |
|---|---|---|---|
| Throughput (tok/s/gpu) | MTP:—Off:931.5 | MTP:—Off:899.5 | MTP:—Off:493.4 |
| Cost ($/M tok) | MTP:—Off:$0.698 | MTP:—Off:$0.723 | MTP:—Off:$1.323 |
| tok/s/MW | MTP:—Off:490276 | MTP:—Off:473406 | MTP:—Off:259700 |
| Concurrency | MTP:—Off:~28 | MTP:—Off:~19 | MTP:—Off:~7 |
Inference Performance
Inference performance metrics across different models, hardware configurations, and serving parameters.