[
    {
        "id": "osp-20640",
        "type": "article-journal",
        "title": "LongSpark: Efficient speculative decoding with a fixed-cost parallel drafter",
        "author": [
            {
                "family": "He",
                "given": "Hao-Yuan"
            },
            {
                "family": "Liu",
                "given": "Pengfei"
            },
            {
                "family": "Shen",
                "given": "Si"
            },
            {
                "family": "Li",
                "given": "Ming"
            }
        ],
        "URL": "https://omanscience.com/ar/articles/longspark-efficient-speculative-decoding-with-a-fixed-cost-parallel-drafter",
        "language": "en",
        "issued": {
            "date-parts": [
                [
                    2026
                ]
            ]
        },
        "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."
    }
]