mlsanityData to 14 Oct 2026

Streaming-dLLM: Accelerating Diffusion LLMs via Suffix Pruning and Dynamic Decoding

arXiv 2601.17917 ↗First posted 25 Jan 2026cs.LG cs.AI cs.CLcited by 7

On the map: Efficient LLM inference and serving → Diffusion LLM and MoE inference

Abstract (arXiv)

Diffusion Large Language Models (dLLMs) offer a compelling paradigm for natural language generation, leveraging parallel decoding and bidirectional attention to achieve superior global coherence compared to autoregressive models. While recent works have accelerated inference via KV cache reuse or heuristic decoding, they overlook the intrinsic inefficiencies within the block-wise diffusion process. Specifically, they suffer from spatial redundancy by modeling informative-sparse suffix regions uniformly and temporal inefficiency by applying fixed denoising schedules across all the decoding process. To address this, we propose Streaming-dLLM, a training-free framework that streamlines inference across both spatial and temporal dimensions. Spatially, we introduce attenuation guided suffix modeling to approximate the full context by pruning redundant mask tokens. Temporally, we employ a dynamic confidence aware strategy with an early exit mechanism, allowing the model to skip unnecessary iterations for converged tokens. Extensive experiments show that Streaming-dLLM achieves up to 68.2X speedup while maintaining generation quality, highlighting its effectiveness in diffusion decoding. The code is available at https://github.com/xiaoshideta/Streaming-dLLM.

Key facts. Cited by 7 papers in the map (data through 14 Oct 2026). Placed in Diffusion LLM and MoE inference, one of 12 directions in Efficient LLM inference and serving. Builds on 42 of its 43 references that are in the map. First posted on arXiv 25 Jan 2026.

Builds on

42 of 43 references.

Closest ancestors · 4

Same direction, newest first

  1. 2025Attention Is All You Need for KV Cache in Diffusion LLMs16 Oct
  2. 2025Sparse-dLLM: Accelerating Diffusion LLMs with Dynamic Cache Eviction4 Aug
  3. 2025Fast-dLLM: Training-free Acceleration of Diffusion LLM by Enabling KV Cache and Parallel Decoding28 May
  4. 2025dKV-Cache: The Cache for Diffusion Language Models21 May

Other references · 38

  1. 2023GPT-4 Technical Reportcited by 18,396
  2. 2024The Llama 3 Herd of Modelscited by 14,715
  3. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  4. 2025Qwen3 Technical Reportcited by 8,820
  5. 2021Evaluating Large Language Models Trained on Codecited by 8,320
  6. 2020Measuring Massive Multitask Language Understandingcited by 7,963
  7. 2021RoFormer: Enhanced Transformer with Rotary Position Embeddingcited by 4,703
  8. 2021Program Synthesis with Large Language Modelscited by 3,637
  9. 2021Structured Denoising Diffusion Models in Discrete State-Spacescited by 1,719
  10. 2022Diffusion-LM Improves Controllable Text Generationcited by 1,027
  11. 2025Large Language Diffusion Modelscited by 912
  12. 2024Simple and Effective Masked Diffusion Language Modelscited by 848
26 more
  1. 2025Dream 7B: Diffusion Large Language Modelscited by 544
  2. 2025Block Diffusion: Interpolating Between Autoregressive and Diffusion Language Modelscited by 409
  3. 2025LLaDA 1.5: Variance-Reduced Preference Optimization for Large Language Diffusion Modelscited by 285
  4. 2025Remasking Discrete Diffusion Models with Inference-Time Scalingcited by 189
  5. 2025d1: Scaling Reasoning in Diffusion Large Language Models via Reinforcement Learningcited by 186
  6. 2025Seed Diffusion: A Large-Scale Diffusion Language Model with High-Speed Inferencecited by 178
  7. 2025dLLM-Cache: Accelerating Diffusion Large Language Models with Adaptive Cachingcited by 169
  8. 2025LLaDA-V: Large Language Diffusion Models with Visual Instruction Tuningcited by 148
  9. 2025Fast-dLLM v2: Efficient Block-Diffusion LLMcited by 139
  10. 2023AR-Diffusion: Auto-Regressive Diffusion Model for Text Generationcited by 118
  11. 2022Text Generation with Diffusion Language Models: A Pre-training Approach with Continuous Paragraph Denoisecited by 79
  12. 2025A Survey on Diffusion Language Modelscited by 78
  13. 2022Latent Diffusion Energy-Based Model for Interpretable Text Modelingcited by 67
  14. 2023TESS: Text-to-Text Self-Conditioned Simplex Diffusioncited by 65
  15. 2025Pangu Embedded: An Efficient Dual-system LLM Reasoner with Metacognitioncited by 62
  16. 2025LLaDA-MoE: A Sparse MoE Diffusion Language Modelcited by 57
  17. 2025Theoretical Benefit and Limitation of Diffusion Language Modelcited by 47
  18. 2025WeDLM: Reconciling Diffusion Language Models with Standard Causal Attention for Fast Inferencecited by 41
  19. 2022Empowering Diffusion Models on the Embedding Space for Text Generationcited by 40
  20. 2025DPad: Efficient Diffusion Language Models with Suffix Dropoutcited by 39
  21. 2025dInfer: An Efficient Inference Framework for Diffusion Language Modelscited by 37
  22. 2023LGViT: Dynamic Early Exiting for Accelerating Vision Transformercited by 31
  23. 2025A Survey on Parallel Text Generation: From Parallel Decoding to Diffusion Language Modelscited by 30
  24. 2025From Next-Token to Next-Block: A Principled Adaptation Path for Diffusion LLMscited by 24
  25. 2025DLM-One: Diffusion Language Models for One-Step Sequence Generationcited by 20
  26. 2025Latent Refinement Decoding: Enhancing Diffusion-Based Language Models by Refining Belief Statescited by 9

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 7.

7 papers.

  1. 2026DMax: Aggressive Parallel Decoding for dLLMscited by 18
  2. 2026TACG: Trajectory-Aware Commit Gating for Diffusion Language Model Decodingcited by 3
  3. 2026Continuous Language Diffusion as a Decoder-Interface Problemcited by 1
  4. 2026Follow the Latent Roadmap: Navigating Revocable Decoding for Diffusion LLMs with Anchor Tokenscited by 1
  5. 2026Accelerating Diffusion Language Models via Structured Suffix Modelingcited by 0
  6. 2026Affix Cache for Diffusion Large Language Modelscited by 0
  7. 2026OLED-MoE: Accelerating MoE-Based dLLM Inference via Inter-Iteration Locality-Aware Expert Offloadingcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Dynamic-dLLM: Dynamic Cache-Budget and Adaptive Parallel Decoding for Training-Free Acceleration of Diffusion LLM
  2. 2026Accelerating Masked Diffusion Large Language Models: A Survey of Efficient Inference Techniques
  3. 2025Orchestrating Dual-Boundaries: An Arithmetic Intensity Inspired Acceleration Framework for Diffusion Language Models
  4. 2026Early-Bird Decoding: Accelerating Diffusion LLMs with Learnable Block Sizes and Parallel Sampling
  5. 2026DAWN: Dependency-Aware Fast Inference for Diffusion LLMs
  6. 2025Diffusion LLMs Can Do Faster-Than-AR Inference via Discrete Diffusion Forcing
  7. 2026Window-Diffusion: Accelerating Diffusion Language Model Inference with Windowed Token Pruning and Caching
  8. 2026Roll Out and Roll Back: Diffusion LLMs are Their Own Efficiency Teachers