SSSD: Simply-Scalable Speculative Decoding

Michele Marzollo, Jiawei Zhuang, Niklas Roemer, Niklas Zwingenberger, Lorenz K Muller, Lukas Cavigelli


Abstract
Speculative Decoding has emerged as a popular technique for accelerating inference in Large Language Models. However, most existing approaches yield only modest improvements in production serving systems. Methods that achieve substantial speedups typically rely on an additional trained draft model or auxiliary model components, increasing deployment and maintenance complexity. This added complexity reduces flexibility, particularly when serving workloads shift to tasks, domains, or languages that are not well represented in the draft model’s training data.We introduce Simply-Scalable Speculative Decoding (SSSD), a training-free method that combines lightweight n-gram matching with hardware-aware speculation. Relative to standard autoregressive decoding, SSSD reduces latency by up to 2.9×. It achieves performance on par with leading training-based approaches across a broad range of benchmarks, while requiring substantially lower adoption effort—no data preparation, training or tuning are needed—and exhibiting superior robustness under language and domain shift, as well as in long-context settings.
Anthology ID:
2026.acl-long.1530
Volume:
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
Month:
July
Year:
2026
Address:
San Diego, California, United States
Editors:
Maria Liakata, Viviane P. Moreira, Jiajun Zhang, David Jurgens
Venue:
ACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
33123–33139
Language:
URL:
https://aclanthology.org/2026.acl-long.1530/
DOI:
10.18653/v1/2026.acl-long.1530
Bibkey:
Cite (ACL):
Michele Marzollo, Jiawei Zhuang, Niklas Roemer, Niklas Zwingenberger, Lorenz K Muller, and Lukas Cavigelli. 2026. SSSD: Simply-Scalable Speculative Decoding. In Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 33123–33139, San Diego, California, United States. Association for Computational Linguistics.
Cite (Informal):
SSSD: Simply-Scalable Speculative Decoding (Marzollo et al., ACL 2026)
Copy Citation:
PDF:
https://aclanthology.org/2026.acl-long.1530.pdf
Checklist:
 2026.acl-long.1530.checklist.pdf