# dUltra: Ultra-Fast Diffusion Language Models via Reinforcement Learning

arXiv: https://arxiv.org/abs/2512.21446 · HTML: https://mlsanity.com/paper/2512.21446/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- Placed in Discrete and text diffusion, one of 14 directions in Diffusion models.
- Builds on 44 of its 48 references that are in the map.
- First posted on arXiv 24 Dec 2025.

## Placed in

- Diffusion models / [Discrete and text diffusion](https://mlsanity.com/diffusion/discrete-and-text-diffusion/)

## Builds on (most cited first, 30 of 44)

- [Denoising Diffusion Probabilistic Models](https://mlsanity.com/paper/2006.11239/) (2020)
- [Structured Denoising Diffusion Models in Discrete State-Spaces](https://mlsanity.com/paper/2107.03006/) (2021)
- [Understanding R1-Zero-Like Training: A Critical Perspective](https://mlsanity.com/paper/2503.20783/) (2025)
- [Qwen2.5-Math Technical Report: Toward Mathematical Expert Model via Self-Improvement](https://mlsanity.com/paper/2409.12122/) (2024)
- [Diffusion-LM Improves Controllable Text Generation](https://mlsanity.com/paper/2205.14217/) (2022)
- [Large Language Diffusion Models](https://mlsanity.com/paper/2502.09992/) (2025)
- [Simple and Effective Masked Diffusion Language Models](https://mlsanity.com/paper/2406.07524/) (2024)
- [Discrete Diffusion Modeling by Estimating the Ratios of the Data Distribution](https://mlsanity.com/paper/2310.16834/) (2023)
- [Simplified and Generalized Masked Diffusion for Discrete Data](https://mlsanity.com/paper/2406.04329/) (2024)
- [Argmax Flows and Multinomial Diffusion: Learning Categorical Distributions](https://mlsanity.com/paper/2102.05379/) (2021)
- [Dream 7B: Diffusion Large Language Models](https://mlsanity.com/paper/2508.15487/) (2025)
- [A Continuous Time Framework for Discrete Denoising Models](https://mlsanity.com/paper/2205.14987/) (2022)
- [DiffuSeq: Sequence to Sequence Text Generation with Diffusion Models](https://mlsanity.com/paper/2210.08933/) (2022)
- [Fast-dLLM: Training-free Acceleration of Diffusion LLM by Enabling KV Cache and Parallel Decoding](https://mlsanity.com/paper/2505.22618/) (2025)
- [Your Absorbing Discrete Diffusion Secretly Models the Conditional Distributions of Clean Data](https://mlsanity.com/paper/2406.03736/) (2024)
- [Mercury: Ultra-Fast Language Models Based on Diffusion](https://mlsanity.com/paper/2506.17298/) (2025)
- [DiffusionBERT: Improving Generative Masked Language Models with Diffusion Models](https://mlsanity.com/paper/2211.15029/) (2022)
- [Train for the Worst, Plan for the Best: Understanding Token Ordering in Masked Diffusions](https://mlsanity.com/paper/2502.06768/) (2025)
- [DiffuCoder: Understanding and Improving Masked Diffusion Models for Code Generation](https://mlsanity.com/paper/2506.20639/) (2025)
- [d1: Scaling Reasoning in Diffusion Large Language Models via Reinforcement Learning](https://mlsanity.com/paper/2504.12216/) (2025)
- [BERT has a Mouth, and It Must Speak: BERT as a Markov Random Field Language Model](https://mlsanity.com/paper/1902.04094/) (2019)
- [dLLM-Cache: Accelerating Diffusion Large Language Models with Adaptive Caching](https://mlsanity.com/paper/2506.06295/) (2025)
- [Likelihood-Based Diffusion Language Models](https://mlsanity.com/paper/2305.18619/) (2023)
- [Concrete Score Matching: Generalized Score Matching for Discrete Data](https://mlsanity.com/paper/2211.00802/) (2022)
- [dKV-Cache: The Cache for Diffusion Language Models](https://mlsanity.com/paper/2505.15781/) (2025)
- [Accelerated Sampling from Masked Diffusion Models via Entropy Bounded Unmasking](https://mlsanity.com/paper/2505.24857/) (2025)
- [Latent Diffusion for Language Generation](https://mlsanity.com/paper/2212.09462/) (2022)
- [Accelerating Diffusion LLMs via Adaptive Parallel Decoding](https://mlsanity.com/paper/2506.00413/) (2025)
- [Revolutionizing Reinforcement Learning Framework for Diffusion Large Language Models](https://mlsanity.com/paper/2509.06949/) (2025)
- [Path Planning for Masked Diffusion Model Sampling](https://mlsanity.com/paper/2502.03540/) (2025)

## Built on by (most cited first, 10 of 10)

- [DMax: Aggressive Parallel Decoding for dLLMs](https://mlsanity.com/paper/2604.08302/) (2026)
- [GDSD: Reinforcement Learning as Guided Denoiser Self-Distillation for Diffusion Language Models](https://mlsanity.com/paper/2605.29398/) (2026)
- [dTRPO: Trajectory Reduction in Policy Optimization of Diffusion Large Language Models](https://mlsanity.com/paper/2603.18806/) (2026)
- [Neither Parallel Nor Sequential: How DiffusionGemma Actually Commits Tokens](https://mlsanity.com/paper/2606.14620/) (2026)
- [Mask-Aware Policy Gradients for Diffusion Language Models](https://mlsanity.com/paper/2607.15200/) (2026)
- [Generalized Discrete Diffusion with Self-Correction](https://mlsanity.com/paper/2603.02230/) (2026)
- [Re-evaluating Confidence Remasking in Masked Diffusion Language Models](https://mlsanity.com/paper/2606.12232/) (2026)
- [From Interface to Inference: Eliciting Any-Order Inference from Any-Order Models](https://mlsanity.com/paper/2607.26504/) (2026)
- [Temporal Self-Distillation: Faster Inference in Discrete Diffusion Language Models](https://mlsanity.com/paper/2609.15177/) (2026)
- [Fork-dLLM: Avoiding the Flexibility Trap in Diffusion Language Models](https://mlsanity.com/paper/2609.39859/) (2026)

## Abstract (arXiv)

Masked diffusion language models (MDLMs) offer the potential for parallel token generation, but most open-source MDLMs decode fewer than 5 tokens per model forward pass even with sophisticated sampling strategies, limiting their parallel generation potential. Existing acceleration methods either rely on fixed confidence-based heuristics or use distillation-based approaches that finetune MDLMs on trajectories generated by a base model, which can become off-policy during finetuning and restrict performance to the quality of the base model's samples. We propose textttdUltra, an on-policy reinforcement learning framework based on Group Relative Policy Optimization (GRPO) that learns unmasking strategies for efficient parallel decoding. dUltra introduces an unmasking planner head that predicts per-token unmasking likelihoods under independent Bernoulli distributions. We jointly optimize the base diffusion LLM and the unmasking order planner using reward signals combining verifiable reward, distillation reward, and the number of unmasking steps. Across mathematical reasoning and code generation tasks, dUltra achieves superior accuracy-efficiency trade-offs compared to state-of-the-art heuristic (Fast-dLLM) and distillation baselines (d3LLM, dParallel), demonstrating that learned unmasking trajectories through on-policy RL enable better exploitation of parallel generation in MDLMs. Code and checkpoints are released at https://github.com/chinsengi/dUltra-os.