mlsanityData to 7 Oct 2026

On the Reasoning Abilities of Masked Diffusion Language Models

arXiv 2510.13117 ↗First posted 15 Oct 2025cs.LG cs.AI cs.CLcited by 6

On the map: LLM reasoning → Diffusion reasoning and test-time scaling

also in 1 more
Abstract (arXiv)

Masked diffusion models (MDMs) for text offer a compelling alternative to traditional autoregressive language models. Parallel generation makes them efficient, but their computational capabilities and the limitations inherent in their parallelism remain largely unexplored. To this end, we characterize what types of reasoning problems MDMs can provably solve and how efficiently. We do this by connecting MDMs to the well-understood reasoning frameworks of chain of thought (CoT) and padded looped transformers (PLTs) in the finite-precision log-width setting: We show that MDMs and polynomially-padded PLTs are, in fact, equivalent in this setting, and that MDMs can solve all problems that CoT-augmented transformers can. Moreover, we showcase classes of problems (including regular languages) for which MDMs are inherently more efficient than CoT transformers, where parallel generation allows for substantially faster reasoning.

Key facts. Cited by 6 papers in the map (data through 7 Oct 2026). Placed in Diffusion reasoning and test-time scaling, one of 17 directions in LLM reasoning. Builds on 30 of its 63 references that are in the map. First posted on arXiv 15 Oct 2025.

Builds on

30 of 63 references.

Closest ancestors · 2

Same direction, newest first

  1. 2024Beyond Autoregression: Discrete Diffusion for Complex Reasoning and Planning18 Oct
  2. 2024Think While You Generate: Discrete Diffusion with Planned Denoising8 Oct

Other references · 28

  1. 2025Large Language Diffusion Modelscited by 872
  2. 2023Discrete Diffusion Modeling by Estimating the Ratios of the Data Distributioncited by 712
  3. 2023Measuring Faithfulness in Chain-of-Thought Reasoningcited by 668
  4. 2018Universal Transformerscited by 613
  5. 2025Fast-dLLM: Training-free Acceleration of Diffusion LLM by Enabling KV Cache and Parallel Decodingcited by 404
  6. 2025Block Diffusion: Interpolating Between Autoregressive and Diffusion Language Modelscited by 385
  7. 2025Reasoning with Latent Thoughts: On the Power of Looped Transformerscited by 223
  8. 2025Train for the Worst, Plan for the Best: Understanding Token Ordering in Masked Diffusionscited by 208
  9. 2025DiffuCoder: Understanding and Improving Masked Diffusion Models for Code Generationcited by 184
  10. 2025Seed Diffusion: A Large-Scale Diffusion Language Model with High-Speed Inferencecited by 173
  11. 2023A Reparameterized Discrete Diffusion Model for Text Generationcited by 152
  12. 2025Accelerated Sampling from Masked Diffusion Models via Entropy Bounded Unmaskingcited by 128
16 more
  1. 2025Mixture-of-Recursions: Learning Dynamic Recursive Depths for Adaptive Token-Level Computationcited by 120
  2. 2025Path Planning for Masked Diffusion Model Samplingcited by 79
  3. 2025A Little Depth Goes a Long Way: The Expressive Power of Log-Depth Transformerscited by 49
  4. 2025Theoretical Benefit and Limitation of Diffusion Language Modelcited by 46
  5. 2024Convergence Analysis of Discrete Diffusion Model: Exact Implementation through Uniformizationcited by 44
  6. 2025A Survey on Parallel Text Generation: From Parallel Decoding to Diffusion Language Modelscited by 27
  7. 2025Breaking AR's Sampling Bottleneck: Provable Acceleration via Diffusion Language Modelscited by 23
  8. 2025Exact Expressive Power of Transformers with Paddingcited by 19
  9. 2025Speed Always Wins: A Survey on Efficient Architectures for Large Language Modelscited by 16
  10. 2025Characterizing the Expressivity of Fixed-Precision Transformer Language Modelscited by 14
  11. 2025The Serial Scaling Hypothesiscited by 11
  12. 2025Unique Hard Attention: A Tale of Two Sidescited by 10
  13. 2025To CoT or To Loop? A Formal Comparison Between Chain-of-Thought and Looped Transformerscited by 10
  14. 2025Pause Tokens Strictly Increase the Expressivity of Constant-Depth Transformerscited by 8
  15. 2023Formal Aspects of Language Modelingcited by 7
  16. 2025Perfect diffusion is mathsfTC0 -- Bad diffusion is Turing-completecited by 3

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 6.

6 papers.

  1. 2026Where-to-Unmask: Ground-Truth-Guided Unmasking Order Learning for Masked Diffusion Language Modelscited by 9
  2. 2026GeoBlock: Inferring Block Granularity from Dependency Geometry in Diffusion Language Modelscited by 1
  3. 2026Revisiting Padded Transformer Expressivity: Which Architectural Choices Matter and Which Don'tcited by 1
  4. 2026Looped Diffusion Language Modelscited by 0
  5. 2026Recursive Scaling in Masked Diffusion Modelscited by 0
  6. 2026Towards Efficient Reasoning: Learning Causal Shortcuts for Diffusion Language Modelscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2024Scaling up Masked Diffusion Models on Text
  2. 2026Parallelism and Generation Order in Masked Diffusion Language Models: Limits Today, Potential Tomorrow
  3. 2025Beyond Surface Reasoning: Unveiling the True Long Chain-of-Thought Capacity of Diffusion Large Language Models
  4. 2025On Powerful Ways to Generate: Autoregression, Diffusion, and Beyond
  5. 2026Mask-Aware Policy Gradients for Diffusion Language Models
  6. 2026Register Tokens for Bounded-State Reasoning in Diffusion Language Models
  7. 2025Coevolutionary Continuous Discrete Diffusion: Make Your Diffusion Language Model a Latent Reasoner
  8. 2026Beyond Masks: Efficient, Flexible Diffusion Language Models via Deletion-Insertion Processes