📄 Paper List

Surveys / Reading Guides

Date	Name	Title	Paper	Github
2025.6	DLLM-Survey	Discrete Diffusion in Large Language and Multimodal Models: A Survey
2025.8	Awesome-DLMs	A Survey of Diffusion Language Models
2026.1	Awesome-Efficient-dLLMs	Efficient Diffusion Language Models: A Comprehensive Survey

---

Paradigm

• space: continuous / discrete
• time: discrete-time / CTMC
• rep: embedding / manifold / token
• noise / mechanism: gaussian / transition-matrix / absorbing-mask / interpolating / block
• extra: masked-lm / conditional / efficient / AR-hybrid / factorization

Date	Name	Title	Paper	Github	Tag
2022.5	Diffusion-LM	Diffusion-LM Improves Controllable Text Generation
2025.2	RDLM	Continuous Diffusion Model for Language Modeling (Riemannian Diffusion LM)
2021.7	D3PM	Structured Denoising Diffusion Models in Discrete State-Spaces		—
2022.5	CTMC-DDM	A Continuous Time Framework for Discrete Denoising Models		—
2023.5	DiffusionBERT	DiffusionBERT: Improving Generative Masked Language Models with Diffusion Models		—
2024.6	ConcreteScore/RADD	Your Absorbing Discrete Diffusion Secretly Models the Conditional Distributions of Clean Data		—
2024.6	MDLM	Simple and Effective Masked Diffusion Language Models
2025.3	GIDD	Generalized Interpolating Discrete Diffusion
2025.3	BD3-LMs	Block Diffusion: Interpolating Between Autoregressive and Diffusion Language Models
2022.11	CDCD	Continuous Diffusion for Categorical Data		—
2022.11	SED	Self-conditioned Embedding Diffusion for Text Generation
2022.10	SSD-LM	SSD-LM: Semi-autoregressive Simplex-based Diffusion Language Model for Text Generation and Modular Control
2023.09	Latent-Diffusion-LM	Latent Diffusion for Language Generation		—
2023.05	Plaid	Likelihood-Based Diffusion Language Models		—
2023.10	SEDD	Discrete Diffusion Modeling by Estimating the Ratios of the Data Distribution (SEDD)
2024.06	MD4	Simplified and Generalized Masked Diffusion for Discrete Data
2024.12	SimpleGuidance	Simple Guidance Mechanisms for Discrete Diffusion Models		—
2025.04	HyperSchedules	Unifying Autoregressive and Diffusion-Based Sequence Generation (DeLTa @ ICLR 2025)		—
2025.09	FlexMDM	Any-Order Flexible Length Masked Diffusion		—
2025.06	DiffusionDuality	The Diffusion Duality (DeLTa @ ICLR 2025)		—
2025.09	Energy-MDM	Masked Diffusion Models as Energy Minimization		—
2026.02	XDLM	Balancing Understanding and Generation in Discrete Diffusion Models
2026.03	GDDS	Generalized Discrete Diffusion from Snapshots

---

Training

Pre-training & Post-training

• focus: scaling-law / base-model / adaptation / instruct
• mechanism: masked / block
• space: discrete / continuous (when relevant)
• domain: general / code
• extra: compute-data-optimal / efficient / noise-type

Date	Name	Title	Paper	Github	Tag
2024.10	SMDM-ScalingLaw	Scaling Up Masked Diffusion Models on Text
2025.10	Quokka	Training Optimal Large Diffusion Language Models
2025.12	DLM-Scaling-Noise	Scaling Behavior of Discrete Diffusion Language Models
2025.02	DLM-Theory	Theoretical Benefit and Limitation of Diffusion Language Model		—
2025.07	DataConstrained	Diffusion Beats Autoregressive in Data-Constrained Settings		—
2023.5	TESS	TESS: Text-to-Text Self-Conditioned Simplex Diffusion
2025.2	TESS-2	TESS 2: A Large-Scale Generalist Diffusion Language Model
2024.10	DiffuLLaMA	Scaling Diffusion Language Models via Adaptation from Autoregressive Models
2025.2	LLaDA	Large Language Diffusion Models (LLaDA)
2025.8	Dream-7B	Dream 7B: Diffusion Large Language Models
2025.9	Dream-Coder-7B	Dream-Coder 7B: An Open Diffusion Language Model for Code
2025.06	DiffuCoder	DiffuCoder: Understanding and Improving Masked Diffusion Models for Code Generation		—	code generation
2025.08	Seed-Diffusion	Seed Diffusion: A Large-Scale Diffusion Language Model with High-Speed Inference
2025.10	SDAR	SDAR: A Synergistic Diffusion-AutoRegression Paradigm for Scalable Sequence Generation
2025.9	Fast-dLLM-v2	Fast-dLLM v2: Efficient Block-Diffusion LLM
2025.12	LLaDA2.0	LLaDA2.0: Scaling Up Diffusion Language Models to 100B
2025.12	NBDiff	From Next-Token to Next-Block: A Principled Adaptation Path for Diffusion LLMs
2025.08	D2F	Diffusion LLMs Can Do Faster-Than-AR Inference via Discrete Diffusion Forcing
2025.10	RND1	RND1: Simple, Scalable AR-to-Diffusion Conversion
2025.09	SDLM	Sequential Diffusion Language Models
2025.09	LLaDA-MoE	LLaDA-MoE: A Sparse MoE Diffusion Language Model		—
2025.12	E2D2	Encoder-Decoder Diffusion Language Models for Efficient Training and Inference
2026.02	LLaDA2.1	LLaDA2.1: Speeding Up Text Diffusion via Token Editing
2026.03	MaskDLLM	Mask Is What DLLM Needs: A Masked Data Training Paradigm for Diffusion LLMs		—

RL / Policy Optimization

• family: GRPO / DPO / PG / boundary-guided / trace
• goal: reasoning / preference / memory-efficient
• target: diffusion-llm / masked-diffusion (when clear)

Date	Name	Title	Paper	Github	Tag
2025.4	D1	d1: Scaling Reasoning in Diffusion Large Language Models via Reinforcement Learning
2025.5	VRPO	LLaDA 1.5: Variance-Reduced Preference Optimization for Large Language Diffusion Models
2025.5	DCoLT	Reinforcing the Diffusion Chain of Lateral Thought with Diffusion Language Models		—
2025.6	Coupled-GRPO	Coupled-GRPO: Enhancing Diffusion Language Models with Coupled Generative Reinforcement Policy Optimization
2025.8	MDPO	MDPO: Overcoming the Training-Inference Divide of Masked Diffusion Language Models
2025.9	TraceRL	TraceRL: Revolutionizing Reinforcement Learning Framework for Diffusion Large Language Models
2025.9	CJ-GRPO	Taming Masked Diffusion Language Models via Consistency Trajectory Reinforcement Learning with Fewer Decoding Step
2025.9	d2	d2: Improved Techniques for Training Reasoning Diffusion Language Models		—
2025.07	wd1	wd1: Weighted Policy Optimization for Reasoning in Diffusion Language Models		—
2025.09	IGPO	Inpainting-Guided Policy Optimization for Diffusion Large Language Models		—
2025.10	SAPO	Step-Aware Policy Optimization for Reasoning in Diffusion Large Language Models
2025.10	BGPO	Boundary-Guided Policy Optimization for Memory-efficient RL of Diffusion Large Language Models
2025.10	SPG	Sandwiched Policy Gradient for Masked Diffusion Language Models
2025.12	ESPO	Principled RL for Diffusion LLMs Emerges from a Sequence-Level Perspective
2025.12	DiRL	DiRL: An Efficient Post-Training Framework for Diffusion Language Models
2026.1	JustGRPO	JustGRPO: Simplifying Generative Reinforcement Policy Optimization for Diffusion Language Models
2026.?	LightningRL	LightningRL: Breaking the Accuracy–Parallelism Trade-off of Block-wise dLLMs via Reinforcement Learning
2026.2	STP	Efficient and Stable Reinforcement Learning for Diffusion Language Models
2026.3	StableDRL	Stabilizing Reinforcement Learning for Diffusion Language Models
2026.3	LFPO	LFPO: Likelihood-Free Policy Optimization for Masked Diffusion Models		—
2025.10	AGRPO	Simple Policy Gradients for Reasoning with Diffusion Language Models
2026.03	EGSPO-SA	Reinforcement Learning for Diffusion LLMs with Entropy-Guided Step Selection and Stepwise Advantages
2026.03	dTRPO	dTRPO: Trajectory Reduction in Policy Optimization of Diffusion Large Language Models

Distillation / Compression

Date	Name	Title	Paper	Github	Tag
2025.08	QDLM	Quantization Meets dLLMs: A Systematic Study of Post-training Quantization for Diffusion LLMs
2025.08	DLLMQuant	DLLMQuant: Quantizing Diffusion-based Large Language Models	—	—
2025.09	Quant-dLLM	Quant-dLLM: Post-Training Extreme Low-Bit Quantization for Diffusion Large Language Models
2025.06	DLM-One	DLM-One: Diffusion Language Models for One-Step Sequence Generation		—
2026.01	CD4LM	CD4LM: Consistency Distillation and aDaptive Decoding for Diffusion Language Models
2026.01	d3LLM	d3LLM: Ultra-Fast Diffusion LLM using Pseudo-Trajectory Distillation
2026.03	D-MMD	Beyond Single Tokens: Distilling Discrete Diffusion Models via Discrete MMD		—

---

Inference & Acceleration

Inference-time Collaboration & Planning

Date	Name	Title	Paper	Github	Tag
2024.02	DoT	Diffusion of Thoughts: Chain-of-Thought Reasoning in Diffusion Language Models (NeurIPS 2024)
2024.10	MDM-Planning	Beyond Autoregression: Discrete Diffusion for Complex Reasoning and Planning (ICLR 2025)
2024.06	SSD-2	David helps Goliath: Inference-Time Collaboration Between Small Specialized and Large General Diffusion LMs (NAACL 2024)		—
2025.06	LongLLaDA	LongLLaDA: Unlocking Long Context Capabilities in Diffusion LLMs
2025.08	Blockwise-SFT	Blockwise SFT for Diffusion Language Models: Reconciling Bidirectional Attention and Autoregressive Decoding		—
2025.11	TiDAR	TiDAR: Think in Diffusion, Talk in Autoregression		—

Efficient Sampling & Decoding

Date	Name	Title	Paper	Github	Tag
2025.05	Fast-dLLM	Fast-dLLM: Training-free Acceleration of Diffusion LLM by Enabling KV Cache and Parallel Decoding
2025.05	EB-Sampler	Accelerated Sampling from Masked Diffusion Models via Entropy Bounded Unmasking		—
2025.05	DINGO	DINGO: Constrained Inference for Diffusion LLMs		—
2025.06	Dilated-Scheduling	Plan for Speed: Dilated Scheduling for Masked Diffusion Language Models		—
2025.06	SlowFast-Sampling	Accelerating diffusion large language models with slowfast: The three golden principles
2025.06	WINO	Wide-In, Narrow-Out: Revokable Decoding for Efficient and Effective DLLMs
2025.06	APD	Accelerating Diffusion LLMs via Adaptive Parallel Decoding
2025.08	Prophet	Diffusion language models know the answer before decoding
2025.08	RWS	Reward-Weighted Sampling: Enhancing Non-Autoregressive Characteristics in Masked Diffusion LLMs	—	—
2025.10	LocalLeap	Accelerated Diffusion LLM Inference via Local Determinism Propagation
2025.10	FreeDave	Free Draft-and-Verification: Toward Lossless Parallel Decoding for Diffusion Large Language Models
2025.10	Saber	Saber: An Efficient Sampling with Adaptive Acceleration and Backtracking Enhanced Remasking for Diffusion Language Model
2025.12	SchED	Fast-Decoding Diffusion Language Models via Progress-Aware Confidence Schedules
2025.12	CadLLM	Improving the Throughput of Diffusion-based Large Language Models via a Training-Free Confidence-Aware Calibration		—
2025.12	LoPA	LoPA: Scaling dLLM Inference via Lookahead Parallel Decoding
2025.09	dParallel	dparallel: Learnable parallel decoding for dllms
2025.09	LSD	Learnable sampler distillation for discrete diffusion models
2025.09	ADJUST	Enabling Approximate Joint Sampling in Diffusion LMs		—
2025.09	Learn2PD	Learning to Parallel: Accelerating Diffusion Large Language Models via Learnable Parallel Decoding
2025.12	Learning-Unmasking-Policies	Learning Unmasking Policies for Diffusion Language Models		—
2025.09	Spiffy	Spiffy: Multiplying Diffusion LLM Acceleration via Lossless Speculative Decoding		—
2025.10	SSD	Self Speculative Decoding for Diffusion Large Language Models		—
2025.10	DiffuSpec	DiffuSpec: Unlocking Diffusion Language Models for Speculative Decoding		—
2025.12	DEER	DEER: Draft with Diffusion, Verify with Autoregressive Models
2026.01	DFlash	DFlash: Block Diffusion for Flash Speculative Decoding	—
2026.01	DART	DART: Diffusion-Inspired Speculative Decoding for Fast LLM Inference
2026.03	ES-dLLM	ES-dLLM: Efficient Inference for Diffusion Large Language Models by Early-Skipping
2026.03	EntropyCache	EntropyCache: Decoded Token Entropy Guided KV Caching for Diffusion Language Models

Other Sampling

---

Multimodal & Applications

Multimodal Understanding

• task: vl-understanding
• route: discrete / block / hybrid / adaptation
• training: vis-instruct (when applicable)

Date	Name	Title	Paper	Github	Tag
2025.5	LLaDA-V	LLaDA-V: Large Language Diffusion Models with Visual Instruction Tuning
2025.5	LaViDa	LaViDa: A Large Diffusion Language Model for Multimodal Understanding
2025.5	Dimple	Dimple: Discrete Diffusion Multimodal Large Language Model with Parallel Decoding
2025.12	SDAR-VL	SDAR-VL: Stable and Efficient Block-wise Diffusion for Vision-Language Understanding
2025.12	DiffusionVL	DiffusionVL: Translating Any Autoregressive Models into Diffusion Vision Language Models
2025.11	D3ToM	D3ToM: Decider-Guided Dynamic Token Merging for Accelerating Diffusion MLLMs
2025.11	Token-Pruning	A Comprehensive Study on Visual Token Redundancy for Discrete Diffusion-based Multimodal Large Language Models		—
2025.12	Sparse-LaViDa	Sparse-LaViDa: Sparse Multimodal Discrete Diffusion Language Models		—

Unified Multimodal Understanding & Generation

• scope: unified (understanding + generation)
• space: discrete (mostly)
• focus: foundation / visual-prior / high-res / omni

Date	Name	Title	Paper	Github	Tag
2025.03	UniDisc	Unified Multimodal Discrete Diffusion
2025.5	MMaDA	MMaDA: Multimodal Large Diffusion Language Models
2025.5	Muddit	Muddit: Liberating Generation Beyond Text-to-Image with a Unified Discrete Diffusion Model
2025.05	FUDOKI	FUDOKI: Discrete Flow-based Unified Understanding and Generation via Kinetic-Optimal Velocities		—
2025.9	LaViDa-O	LaViDa-O: Elastic Masked Diffusion Models for Unified Multimodal Understanding and Generation
2025.10	Lumina-DiMOO	Lumina-DiMOO: An Omni Diffusion Large Language Model for Multi-Modal Generation and Understanding
2026.03	LLaDA-o	LLaDA-o: An Effective and Length-Adaptive Omni Diffusion Model
2026.03	Omni-Diffusion	Omni-Diffusion: Unified Multimodal Understanding and Generation with Masked Discrete Diffusion

Applications

Date	Name	Title	Paper	Github	Tag
2025.08	DD-VLA	Discrete Diffusion VLA: Bringing Discrete Diffusion to Action Decoding in Vision-Language-Action Policies		—
2025.09	LLaDA-VLA	LLaDA-VLA: Vision Language Diffusion Action Models		—
2026.03	MinerU-Diffusion	MinerU-Diffusion: Rethinking Document OCR as Inverse Rendering via Diffusion Decoding

Others

Date	Name	Title	Paper	Github	Note
2025.08	Thinking-Inside-the-Mask	Thinking Inside the Mask: In-Place Prompting in Diffusion LLMs		—	prompting / in-place prompting
2025.08	Time-Is-a-Feature	Time Is a Feature: Exploiting Temporal Dynamics in Diffusion Language Models		—	temporal dynamics
2025.09	A2D	A2D: Any-Order, Any-Step Safety Alignment for Diffusion Language Models		—	safety alignment
2025.07	Jailbreaking-LDLMs	Jailbreaking Large Language Diffusion Models: Revealing Hidden Safety Flaws in Diffusion-Based Text Generation		—	safety attack
2025.07	Devil-behind-the-mask	The Devil behind the mask: An emergent safety vulnerability of Diffusion LLMs		—	safety vulnerability
2025.10	ParallelBench	ParallelBench: Understanding the Trade-offs of Parallel Decoding in Diffusion LLMs		—
2026.03	Thinking-EoS-by-EoS	Diffusion LLMs can think EoS-by-EoS			hidden scratchpad / EoS tokens / reasoning
2026.03	MemDLM	MemDLM: Memory-Enhanced DLM Training			memory-enhanced training / train-inference mismatch / in-weight retrieval
2026.02	Latent-Tokens	Reasoning with Latent Tokens in Diffusion Language Models		—	latent tokens / reasoning / speed-quality trade-off