# Accelerating Large-Scale Reasoning Model Inference with Sparse Self-Speculative Decoding

arXiv: https://arxiv.org/abs/2512.01278 · HTML: https://mlsanity.com/paper/2512.01278/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning.
- Builds on 38 of its 47 references that are in the map.
- First posted on arXiv 1 Dec 2025.

## Placed in

- LLM reasoning / [Test-time compute and efficient thinking](https://mlsanity.com/lr/test-time-compute-and-efficient-thinking/)
- Efficient LLM inference and serving / [Speculative decoding](https://mlsanity.com/inference/speculative-decoding/)

## Builds on (most cited first, 30 of 38)

- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [Efficient Memory Management for Large Language Model Serving with PagedAttention](https://mlsanity.com/paper/2309.06180/) (2023)
- [Efficient Streaming Language Models with Attention Sinks](https://mlsanity.com/paper/2309.17453/) (2023)
- [LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code](https://mlsanity.com/paper/2403.07974/) (2024)
- [Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters](https://mlsanity.com/paper/2408.03314/) (2024)
- [Fast Inference from Transformers via Speculative Decoding](https://mlsanity.com/paper/2211.17192/) (2022)
- [OpenAI o1 System Card](https://mlsanity.com/paper/2412.16720/) (2024)
- [Kimi k1.5: Scaling Reinforcement Learning with LLMs](https://mlsanity.com/paper/2501.12599/) (2025)
- [SGLang: Efficient Execution of Structured Language Model Programs](https://mlsanity.com/paper/2312.07104/) (2023)
- [OlympiadBench: A Challenging Benchmark for Promoting AGI with Olympiad-Level Bilingual Multimodal Scientific Problems](https://mlsanity.com/paper/2402.14008/) (2024)
- [Accelerating Large Language Model Decoding with Speculative Sampling](https://mlsanity.com/paper/2302.01318/) (2023)
- [SnapKV: LLM Knows What You are Looking for Before Generation](https://mlsanity.com/paper/2404.14469/) (2024)
- [H2O: Heavy-Hitter Oracle for Efficient Generative Inference of Large Language Models](https://mlsanity.com/paper/2306.14048/) (2023)
- [DeepSeek-V3 Technical Report](https://mlsanity.com/paper/2412.19437/) (2024)
- [EAGLE: Speculative Sampling Requires Rethinking Feature Uncertainty](https://mlsanity.com/paper/2401.15077/) (2024)
- [Quest: Query-Aware Sparsity for Efficient Long-Context LLM Inference](https://mlsanity.com/paper/2406.10774/) (2024)
- [EAGLE-2: Faster Inference of Language Models with Dynamic Draft Trees](https://mlsanity.com/paper/2406.16858/) (2024)
- [EAGLE-3: Scaling up Inference Acceleration of Large Language Models via Training-Time Test](https://mlsanity.com/paper/2503.01840/) (2025)
- [rStar-Math: Small LLMs Can Master Math Reasoning with Self-Evolved Deep Thinking](https://mlsanity.com/paper/2501.04519/) (2025)
- [Break the Sequential Dependency of LLM Inference Using Lookahead Decoding](https://mlsanity.com/paper/2402.02057/) (2024)
- [Atom: Low-bit Quantization for Efficient and Accurate LLM Serving](https://mlsanity.com/paper/2310.19102/) (2023)
- [FlashInfer: Efficient and Customizable Attention Engine for LLM Inference Serving](https://mlsanity.com/paper/2501.01005/) (2025)
- [InfLLM: Training-Free Long-Context Extrapolation for LLMs with an Efficient Context Memory](https://mlsanity.com/paper/2402.04617/) (2024)
- [Hydra: Sequentially-Dependent Draft Heads for Medusa Decoding](https://mlsanity.com/paper/2402.05109/) (2024)
- [Online Speculative Decoding](https://mlsanity.com/paper/2310.07177/) (2023)
- [TriForce: Lossless Acceleration of Long Sequence Generation with Hierarchical Speculative Decoding](https://mlsanity.com/paper/2404.11912/) (2024)
- [PQCache: Product Quantization-based KVCache for Long Context LLM Inference](https://mlsanity.com/paper/2407.12820/) (2024)
- [MagicDec: Breaking the Latency-Throughput Tradeoff for Long Context Generation with Speculative Decoding](https://mlsanity.com/paper/2408.11049/) (2024)
- [NanoFlow: Towards Optimal Large Language Model Serving Throughput](https://mlsanity.com/paper/2408.12757/) (2024)
- [Twilight: Adaptive Attention Sparsity with Hierarchical Top-p Pruning](https://mlsanity.com/paper/2502.02770/) (2025)

## Built on by (most cited first, 10 of 10)

- [LLM-42: Enabling Determinism in LLM Inference with Verified Speculation](https://mlsanity.com/paper/2601.17768/) (2026)
- [Sparse-RL: Breaking the Memory Wall in LLM Reinforcement Learning via Stable Sparse Rollouts](https://mlsanity.com/paper/2601.10079/) (2026)
- [Vegas: Self-Speculative Decoding with Verification-Guided Sparse Attention](https://mlsanity.com/paper/2602.07223/) (2026)
- [DFlow: Enabling Verifier Information Flow in Block Diffusion Speculative Decoding](https://mlsanity.com/paper/2609.06498/) (2026)
- [A Sparse Glimpse of the Whole: Train-Free Self-Speculative Decoding](https://mlsanity.com/paper/2607.27735/) (2026)
- [ASPIRE: Asynchronous Batched Self-Speculative Decoding for Long-Context LLM Inference](https://mlsanity.com/paper/2609.17943/) (2026)
- [TreeSpark: Calibrated, Load-Adaptive Draft Trees for Semi-Autoregressive Speculative Decoding](https://mlsanity.com/paper/2609.22098/) (2026)
- [Self-Speculation for Faster Reasoning Models](https://mlsanity.com/paper/2608.20359/) (2026)
- [Where Activation Sparsity and KV-Cache Sparsity Cross in LLM Decoding](https://mlsanity.com/paper/2609.33889/) (2026)
- [S3R: Fast and Faithful RL Rollouts via Self-Sparse Speculative Decoding](https://mlsanity.com/paper/2610.13202/) (2026)

## Abstract (arXiv)

Reasoning language models have demonstrated remarkable capabilities on challenging tasks by generating elaborate chain-of-thought (CoT) solutions. However, such lengthy generation shifts the inference bottleneck from compute-bound to memory-bound. To generate each token, the model applies full attention to all previously generated tokens, requiring memory access to an increasingly large KV-Cache. Consequently, longer generations demand more memory access for every step, leading to substantial pressure on memory bandwidth. To address this, we introduce SparseSpec, a speculative decoding framework that reuses the same model as the draft and target models (i.e., self-speculation). SparseSpec features a novel sparse attention mechanism, PillarAttn, as the draft model, which accurately selects critical tokens via elegantly reusing information from the verification stage. Furthermore, SparseSpec co-designs self-speculation with three system innovations: (1) a unified scheduler to batch token drafting and verification, (2) delayed verification for CPU/GPU overlap, and (3) dynamic KV-Cache management to maximize memory utilization. Across various models and datasets, SparseSpec outperforms state-of-the-art solutions, with an up to 2.13x throughput speedup.