OoO-Spec cuts tool-call latency by having a portable 0.6B sidecar draft the whole call's semantics in parallel, reaching 3.89x mean speedup over autoregressive decoding across seven targets and three benchmarks.
Proceedings of the 40th International Conference on Machine Learning , year =
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
OoO-Spec: Out-of-Order Semantic Speculation for Fast Tool Calling
OoO-Spec cuts tool-call latency by having a portable 0.6B sidecar draft the whole call's semantics in parallel, reaching 3.89x mean speedup over autoregressive decoding across seven targets and three benchmarks.