# Diffusion LLM with Native Variable Generation Lengths: Let [EOS] Lead the Way

arXiv: https://arxiv.org/abs/2510.24605 · HTML: https://mlsanity.com/paper/2510.24605/

## Key facts

- Cited by 22 papers in the map (data through 14 Oct 2026).
- Builds on 18 of its 21 references that are in the map.
- First posted on arXiv 28 Oct 2025.

## Builds on (most cited first, 18 of 18)

- [Fast-dLLM: Training-free Acceleration of Diffusion LLM by Enabling KV Cache and Parallel Decoding](https://mlsanity.com/paper/2505.22618/) (2025)
- [Block Diffusion: Interpolating Between Autoregressive and Diffusion Language Models](https://mlsanity.com/paper/2503.09573/) (2025)
- [Better & Faster Large Language Models via Multi-token Prediction](https://mlsanity.com/paper/2404.19737/) (2024)
- [LLaDA 1.5: Variance-Reduced Preference Optimization for Large Language Diffusion Models](https://mlsanity.com/paper/2505.19223/) (2025)
- [Remasking Discrete Diffusion Models with Inference-Time Scaling](https://mlsanity.com/paper/2503.00307/) (2025)
- [DiffuCoder: Understanding and Improving Masked Diffusion Models for Code Generation](https://mlsanity.com/paper/2506.20639/) (2025)
- [Seed Diffusion: A Large-Scale Diffusion Language Model with High-Speed Inference](https://mlsanity.com/paper/2508.02193/) (2025)
- [dLLM-Cache: Accelerating Diffusion Large Language Models with Adaptive Caching](https://mlsanity.com/paper/2506.06295/) (2025)
- [LLaDA-V: Large Language Diffusion Models with Visual Instruction Tuning](https://mlsanity.com/paper/2505.16933/) (2025)
- [Fast-dLLM v2: Efficient Block-Diffusion LLM](https://mlsanity.com/paper/2509.26328/) (2025)
- [SDAR: A Synergistic Diffusion-AutoRegression Paradigm for Scalable Sequence Generation](https://mlsanity.com/paper/2510.06303/) (2025)
- [Diffusion LLMs Can Do Faster-Than-AR Inference via Discrete Diffusion Forcing](https://mlsanity.com/paper/2508.09192/) (2025)
- [Dream-Coder 7B: An Open Diffusion Language Model for Code](https://mlsanity.com/paper/2509.01142/) (2025)
- [dParallel: Learnable Parallel Decoding for dLLMs](https://mlsanity.com/paper/2509.26488/) (2025)
- [Any-Order Flexible Length Masked Diffusion](https://mlsanity.com/paper/2509.01025/) (2025)
- [Don't Settle Too Early: Self-Reflective Remasking for Diffusion Language Models](https://mlsanity.com/paper/2509.23653/) (2025)
- [dInfer: An Efficient Inference Framework for Diffusion Language Models](https://mlsanity.com/paper/2510.08666/) (2025)
- [d2Cache: Accelerating Diffusion-Based LLMs via Dual Adaptive Caching](https://mlsanity.com/paper/2509.23094/) (2025)

## Built on by (most cited first, 22 of 22)

- [Improved Large Language Diffusion Models](https://mlsanity.com/paper/2606.25331/) (2026)
- [SDAR-VL: Stable and Efficient Block-wise Diffusion for Vision-Language Understanding](https://mlsanity.com/paper/2512.14068/) (2025)
- [DAWN: Dependency-Aware Fast Inference for Diffusion LLMs](https://mlsanity.com/paper/2602.06953/) (2026)
- [LLaDA-o: An Effective and Length-Adaptive Omni Diffusion Model](https://mlsanity.com/paper/2603.01068/) (2026)
- [ρ-textttEOS: Training-free Bidirectional Variable-Length Control for Masked Diffusion LLMs](https://mlsanity.com/paper/2601.22527/) (2026)
- [DSB: Dynamic Sliding Block Scheduling for Diffusion LLMs](https://mlsanity.com/paper/2602.05992/) (2026)
- [When to Commit? Towards Variable-Size Self-Contained Blocks for Discrete Diffusion Language Models](https://mlsanity.com/paper/2604.23994/) (2026)
- [The Bitter Lesson of Diffusion Language Models for Agentic Workflows: A Comprehensive Reality Check](https://mlsanity.com/paper/2601.12979/) (2026)
- [Orchestrating Dual-Boundaries: An Arithmetic Intensity Inspired Acceleration Framework for Diffusion Language Models](https://mlsanity.com/paper/2511.21759/) (2025)
- [Diffusion Language Models Are Natively Length-Aware](https://mlsanity.com/paper/2603.06123/) (2026)
- [Seeing the End at Step Zero: Accelerating Diffusion MLLMs via MLP Sparsity-Aware Truncation](https://mlsanity.com/paper/2607.14557/) (2026)
- [Improving Variable-Length Generation in Diffusion Language Models via Length Regularization](https://mlsanity.com/paper/2602.07546/) (2026)
- [Search or Accelerate: Confidence-Switched Position Beam Search for Diffusion Language Models](https://mlsanity.com/paper/2602.10953/) (2026)
- [Answer First, Reason Later: When Commitment Order Costs Accuracy in Diffusion Language Models](https://mlsanity.com/paper/2608.05687/) (2026)
- [Plan, Verify and Fill: A Structured Parallel Decoding Approach for Diffusion Language Models](https://mlsanity.com/paper/2601.12247/) (2026)
- [DOS: Dependency-Oriented Sampler for Masked Diffusion Language Models](https://mlsanity.com/paper/2603.15340/) (2026)
- [VoidPadding: Let [VOID] Handle Padding in Masked Diffusion Language Models so that [EOS] Can Focus on Semantic Termination](https://mlsanity.com/paper/2606.17999/) (2026)
- [Reconsidering Positional Supervision in Masked Diffusion Language Model Training](https://mlsanity.com/paper/2601.22947/) (2026)
- [Diffusion Large Language Models for Visual Speech Recognition](https://mlsanity.com/paper/2605.28456/) (2026)
- [Length-Adaptive Decoding for Masked Diffusion Machine Translation](https://mlsanity.com/paper/2608.22274/) (2026)
- [DRRG: A Discrete Diffusion Framework for Radiology Report Generation](https://mlsanity.com/paper/2608.24105/) (2026)
- [Routing by Reasoning Need: Trajectory-Aware Decoding Control for Diffusion Vision-Language Models](https://mlsanity.com/paper/2609.11315/) (2026)

## Abstract (arXiv)

Diffusion-based large language models (dLLMs) have exhibited substantial potential for parallel text generation, which may enable more efficient generation compared to autoregressive models. However, current dLLMs suffer from fixed generation lengths, which indicates the generation lengths of dLLMs have to be determined before decoding as a hyper-parameter, leading to issues in efficiency and flexibility. To solve these problems, in this work, we propose to train a diffusion LLM with native variable generation lengths, abbreviated as dLLM-Var. Concretely, we aim to train a model to accurately predict the [EOS] token in the generated text, which makes a dLLM be able to natively infer in a block diffusion manner, while still maintaining the ability of global bi-directional (full) attention and high parallelism. Experiments on standard benchmarks demonstrate that our method achieves a 30.1x speedup over traditional dLLM inference paradigms and a 2.4x speedup relative to autoregressive models such as Qwen and Llama. Our method achieves higher accuracy and faster inference, elevating dLLMs beyond mere academic novelty and supporting their practical use in real-world applications. Codes and models have been released.