Plan-and-Patch: Diffusion Language Models for Agentic Planning
cs.AI, cs.CL, cs.LG
Submitted: 2026-10-07
Updated: 2026-10-07
Code: https://github.com/inclusionAI/LLaDA2.X
License: http://creativecommons.org/licenses/by/4.0/
The gist: Planning is increasingly important for long-horizon agents, where successful execution requires coordinating subgoals, tool use, and intermediate outcomes over many steps.
Terminology
Abstract
Planning is increasingly important for long-horizon agents, where successful execution requires coordinating subgoals, tool use, and intermediate outcomes over many steps. Yet assumptions made during planning may be invalidated by the environment, tools may return unexpected results, or actions may fail. Effective agents must therefore not only generate plans, but also revise them. Such revisions often affect only part of a plan, leaving the preceding and subsequent structure intact. Rather than regenerate the entire plan and risk unnecessary changes, repair can regenerate the affected region conditioned on the preserved prefix and suffix. We introduce Plan-and-Patch, a plan-and-act framework in which a diffusion language model (dLLM) generates a structured, program-like plan through parallel unmasking and repairs it by filling in selected regions while keeping the surrounding steps fixed. We compare DreamReasoner-8B and Qwen3-8B as diffusion and autoregressive (AR) planners. On Natural Plan without task-specific training, diffusion (53.7%) achieves nearly twice the plan repair success rate of AR (27.0%). After task-specific training on agentic benchmarks, ALFWorld and TextCraft, the planners achieve similar observed success in plan generation, while diffusion reduces mean plan-generation latency by 39-46% relative to AR. Our results show that Plan-and-Patch provides a framework for faster plan generation and effective plan repair in long-horizon agents.
Sources
- From Output to Evaluation: Does Raw Instruction-Tuned Code LLMs Output Suffice for Fill-in-the-Middle Code Generation?
- Is Conditional Generative Modeling all you need for Decision-Making?
- Structured Denoising Diffusion Models in Discrete State-Spaces
- Efficient Training of Language Models to Fill in the Middle
- Latent-DARM: Bridging Discrete Diffusion And Autoregressive Models For Reasoning
- Masked Diffusion Language Models are Strong and Steerable Text-Based World Models for Agentic RL
- Plan-and-Act: Improving Planning of Agents for Long-Horizon Tasks
- InCoder: A Generative Model for Code Infilling and Synthesis
- DiffuCoder: Understanding and Improving Masked Diffusion Models for Code Generation
- ContextFlow: Hierarchical Task-State Alignment for Long-Horizon Embodied Agents
- Mercury: Ultra-Fast Language Models Based on Diffusion
- Planning with Diffusion for Flexible Behavior Synthesis
- Beyond Next-Token Prediction: A Performance Characterization of Diffusion versus Autoregressive Language Models
- Localizing and Correcting Errors for LLM-based Planners
- Code as Policies: Language Model Programs for Embodied Control
- REFLECT: Summarizing Robot Experiences for Failure Explanation and Correction
- Discrete Diffusion Modeling by Estimating the Ratios of the Data Distribution
- The Bitter Lesson of Diffusion Language Models for Agentic Workflows: A Comprehensive Reality Check
- Repair or Resample? Rethinking Failure Debugging in LLM Multi-Agent Systems
- Diffusion In Diffusion: Reclaiming Global Coherence in Semi-Autoregressive Diffusion
Related papers
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Agents' Last Exam
- The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
- MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
- Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems