Abstract
Speculative decoding accelerates autoregressive inference by verifying multiple draft tokens in a single target forward pass. However, as the context grows, existing state-of-the-art drafters become increasingly expensive, eroding the very efficiency advantage they are designed to provide. We argue that this scaling is unnecessary. A standalone language model must grow with its prefix because it is solely responsible for every token it produces. A drafter, by contrast, only proposes candidates; the target catches and corrects every error before any token is committed. The drafter's decoding cost can therefore be made entirely independent of the prefix length. We introduce LongSpark, a block-diffusion drafter that achieves this by extracting fixed-size, multiscale views from the target's verification pass, thereby eliminating the need for a growing persistent state. Extensive evaluations demonstrate that LongSpark achieves state-of-the-art end-to-end efficiency across multiple model scales and realistic serving conditions. Notably, it delivers the lowest time-per-output-token on long-context tasks while reducing the drafter's context state by several orders of magnitude.
Keywords
Publication details
- Journal
- Not available
- Open access
- Green open access
Cite this article
APA 7
He, H. Y., Liu, P., Shen, S., & Li, M. (2026). LongSpark: Efficient speculative decoding with a fixed-cost parallel drafter. https://omanscience.com/en/articles/longspark-efficient-speculative-decoding-with-a-fixed-cost-parallel-drafter
MLA 9
He, Hao-Yuan, et al. "LongSpark: Efficient speculative decoding with a fixed-cost parallel drafter." https://omanscience.com/en/articles/longspark-efficient-speculative-decoding-with-a-fixed-cost-parallel-drafter.
Chicago (author–date)
He, Hao-Yuan, Pengfei Liu, Si Shen, and Ming Li. 2026. "LongSpark: Efficient speculative decoding with a fixed-cost parallel drafter." https://omanscience.com/en/articles/longspark-efficient-speculative-decoding-with-a-fixed-cost-parallel-drafter.
Harvard
He, H. Y., Liu, P., Shen, S. and Li, M. (2026) 'LongSpark: Efficient speculative decoding with a fixed-cost parallel drafter', Available at: https://omanscience.com/en/articles/longspark-efficient-speculative-decoding-with-a-fixed-cost-parallel-drafter.
Vancouver
He HY, Liu P, Shen S, Li M. LongSpark: Efficient speculative decoding with a fixed-cost parallel drafter. https://omanscience.com/en/articles/longspark-efficient-speculative-decoding-with-a-fixed-cost-parallel-drafter
IEEE
H. Y. He, P. Liu, S. Shen, and M. Li, "LongSpark: Efficient speculative decoding with a fixed-cost parallel drafter," https://omanscience.com/en/articles/longspark-efficient-speculative-decoding-with-a-fixed-cost-parallel-drafter.