DeepSeekv4 Pro 0813 1.6T · Speculative Decoding

MI300X FP8: MTP vs Off Speculative Decoding

Speculative decoding comparison of MTP versus Off on MI300X FP8 (AMD CDNA 3) running DeepSeekv4 Pro 0813 1.6T. 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 comparability

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.

MTP posts 370 tok/s/chip for $0.71 per million tokens at 9 tok/s/user on DeepSeekv4 Pro 0813 1.6T (MI300X FP8); Off posts 243 tok/s/chip for $1.09. MTP is 52% cheaper per token; MTP delivers 52% more tok/s/chip. Draft-token acceptance rates determine whether speculative decoding helps or hurts at a given concurrency level.

Throughput at 13 tok/s/user on DeepSeekv4 Pro 0813 1.6T (MI300X FP8): MTP hits 263 tok/s/chip, Off hits 135. Per-million costs land at $1.01 and $1.96 respectively. MTP is 95% cheaper per token; MTP delivers 95% 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.

Toward the upper edge of the 5–20 tok/s/user interactivity band, at 17 tok/s/user on DeepSeekv4 Pro 0813 1.6T (MI300X FP8): MTP runs 163 tok/s/chip at $1.62/M tokens, Off runs 67 at $3.94/M. MTP is 143% cheaper per token; MTP delivers 143% more tok/s/chip. Gains from speculative decoding vary by workload; short-output prompts tend to benefit less. (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.)

DeepSeekv4 Pro 0813 1.6T: MI300X FP8 MTP versus Off speculative decoding comparison at matched interactivity levels
MI300X FP8 MTP versus Off speculative decoding comparison for this page's canonical default workload.
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)
MTP:370.4Off:243.1
MTP:262.5Off:134.6
MTP:162.9Off:67.0
Cost ($/M tok)
MTP:$0.713Off:$1.085
MTP:$1.005Off:$1.961
MTP:$1.620Off:$3.936
tok/s/MW
MTP:266449Off:174897
MTP:188861Off:96819
MTP:117202Off:48233
Concurrency
MTP:~170Off:~109
MTP:~84Off:~42
MTP:~39Off:~16

Inference Performance

Agentic inference metrics from the AgentX scenario and fixed-sequence inference metrics across models, hardware configurations, and serving parameters.