mlsanityData to 14 Oct 2026

SpecDiff-2: Scaling Diffusion Drafter Alignment For Faster Speculative Decoding

arXiv 2511.00606 ↗First posted 1 Nov 2025cs.CLcited by 26

On the map: Diffusion models → Discrete and text diffusion

also in 1 more
Abstract (arXiv)

Speculative decoding has become the standard approach for accelerating Large Language Model (LLM) inference. It exploits a lossless draft-then-verify procedure to circumvent the latency of autoregressive decoding, achieving impressive speed-ups. Yet, current speculative decoding approaches remain limited by two fundamental bottlenecks: (1) the autoregressive dependency during drafting which limits parallelism, and (2) frequent rejections of draft tokens caused by misalignment between the draft and verify models. This paper proposes SpecDiff-2, a novel framework to jointly address these two bottlenecks. It leverages discrete diffusion as a non-autoregressive drafter to address bottleneck (1) and develops novel techniques to calibrate discrete diffusion drafters with autoregressive verifiers, addressing bottleneck (2). Experimental results across a comprehensive benchmark suite show that SpecDiff-2 achieves a new state-of-the-art across reasoning, coding, and mathematical benchmarks, improving tokens-per-second by up to an average of +55% over previous baselines and obtaining up to 5.5x average speed-up over standard decoding, without any loss of accuracy.

Key facts. Cited by 26 papers in the map (data through 14 Oct 2026). Placed in Discrete and text diffusion, one of 14 directions in Diffusion models. Builds on 28 of its 29 references that are in the map. First posted on arXiv 1 Nov 2025.

Builds on

28 of 29 references.

Closest ancestors · 4

Same direction, newest first

  1. 2025DiffuCoder: Understanding and Improving Masked Diffusion Models for Code Generation25 Jun
  2. 2024Scaling Diffusion Language Models via Adaptation from Autoregressive Models23 Oct
  3. 2024Simple and Effective Masked Diffusion Language Models11 Jun
  4. 2024Simplified and Generalized Masked Diffusion for Discrete Data6 Jun

Other references · 24

  1. 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 13,028
  2. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
  3. 2021Evaluating Large Language Models Trained on Codecited by 8,320
  4. 2022Self-Consistency Improves Chain of Thought Reasoning in Language Modelscited by 5,708
  5. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,371
  6. 2023Toolformer: Language Models Can Teach Themselves to Use Toolscited by 4,182
  7. 2023GPQA: A Graduate-Level Google-Proof Q&A Benchmarkcited by 2,879
  8. 2022GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformerscited by 1,919
  9. 2022Fast Inference from Transformers via Speculative Decodingcited by 1,653
  10. 2023Accelerating Large Language Model Decoding with Speculative Samplingcited by 1,006
  11. 2022STaR: Bootstrapping Reasoning With Reasoningcited by 940
  12. 2023A Simple and Effective Pruning Approach for Large Language Modelscited by 800
12 more
  1. 2024EAGLE: Speculative Sampling Requires Rethinking Feature Uncertaintycited by 578
  2. 2024EAGLE-2: Faster Inference of Language Models with Dynamic Draft Treescited by 370
  3. 2025EAGLE-3: Scaling up Inference Acceleration of Large Language Models via Training-Time Testcited by 331
  4. 2025A Survey on Test-Time Scaling in Large Language Models: What, How, Where, and How Well?cited by 223
  5. 2023DistillSpec: Improving Speculative Decoding via Knowledge Distillationcited by 165
  6. 2023SpecTr: Fast Speculative Decoding via Optimal Transportcited by 129
  7. 2024Hydra: Sequentially-Dependent Draft Heads for Medusa Decodingcited by 117
  8. 2024Speculative Diffusion Decoding: Accelerating Language Generation through Diffusioncited by 57
  9. 2025Think Twice: Enhancing LLM Reasoning by Scaling Multi-round Test-time Thinkingcited by 46
  10. 2024Decoding Speculative Decodingcited by 34
  11. 2024DySpec: Faster Speculative Decoding with Dynamic Token Tree Structurecited by 25
  12. 2025Overclocking LLM Reasoning: Monitoring and Controlling Thinking Path Lengths in LLMscited by 22

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 1; 2026: 25.

26 papers.

Sort:

By area:

  1. 2026DFlash: Block Diffusion for Flash Speculative Decodingcited by 135
  2. 2026DART: Diffusion-Inspired Speculative Decoding for Fast LLM Inferencecited by 24
  3. 2026LK Losses: Direct Acceptance Rate Optimization for Speculative Decodingcited by 18
  4. 2026D-PACE: Dynamic Position-Aware Cross-Entropy for Parallel Speculative Draftingcited by 17
  5. 2026PARD-2: Target-Aligned Parallel Draft Model for Dual-Mode Speculative Decodingcited by 8
  6. 2026xPress: Parallel Refinement for Diffusion Drafters in Speculative Decodingcited by 8
  7. 2026Verification-Aware Training for Speculative Decodingcited by 5
  8. 2025Fail Fast, Win Big: Rethinking the Drafting Strategy in Speculative Decoding via Diffusion LLMscited by 4
  9. 2026TreeFlash: Parallel AR-Approximation for Faster Speculative Decodingcited by 4
  10. 2026DeLS-Spec: Decoupled Long-Short Contexts for Parallel Speculative Draftingcited by 4
  11. 2026Teaching Diffusion to Speculate Left-to-Rightcited by 3
  12. 2026Spec-AUF: Accept-Until-Fail Training under Train-Inference Misalignment for Masked Block Drafterscited by 3
14 more
  1. 2026PRESTO: Prefix-Aligned Tree Drafting for Diffusion Speculative Decodingcited by 3
  2. 2026SimSD: Simple Speculative Decoding in Diffusion Language Modelscited by 2
  3. 2026BlockPilot: Instance-Adaptive Policy Learning for Diffusion-based Speculative Decodingcited by 2
  4. 2026Unlocking Lossless Speedups in LLMs via Discrete Diffusioncited by 2
  5. 2026DFlow: Enabling Verifier Information Flow in Block Diffusion Speculative Decodingcited by 2
  6. 2026TreeSpark: Calibrated, Load-Adaptive Draft Trees for Semi-Autoregressive Speculative Decodingcited by 1
  7. 2026Fully Rotation-Equivariant Spectral-Spatial Learning for Multispectral Object Detectioncited by 0
  8. 2026LibraSpec: Dynamic Diffusion-Based Speculative Decoding via Marginal-Gain-Driven Optimizationcited by 0
  9. 2026From Positionwise Confidence to Prefix Scheduling: Verifier Skipping in Speculative Decodingcited by 0
  10. 2026GRAFT: Adaptive DLM-Based Draft Tree Construction with Target-Distilled Edge Scoringcited by 0
  11. 2026Diffusion Drafts, AR Verifies: Accelerating Document OCR with Self-Speculative Decodingcited by 0
  12. 2026OLED-MoE: Accelerating MoE-Based dLLM Inference via Inter-Iteration Locality-Aware Expert Offloadingcited by 0
  13. 2026DEdit: Iterative Draft Editing for Speculative Decodingcited by 0
  14. 2026BitNest: Bit-Nested Speculative Decoding for Memory-Efficient LLM Inference Accelerationcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026D^2SD: Accelerating Speculative Decoding with Dual Diffusion Draft Models
  2. 2025DiffuSpec: Unlocking Diffusion Language Models for Speculative Decoding
  3. 2026LongSpark: Efficient speculative decoding with a fixed-cost parallel drafter
  4. 2026AdaFlash: Adaptive Speculative Decoding via On-Policy Distilled Diffusion Drafters
  5. 2025Self Speculative Decoding for Diffusion Large Language Models
  6. 2026Accepted Prefixes Are Not All You Need: A Negative Result on PEFT-Based Block-Diffusion Drafting
  7. 2025Draft, Verify, and Improve: Toward Training-Aware Speculative Decoding
  8. 2024Closer Look at Efficient Inference Methods: A Survey of Speculative Decoding