mlsanityData to 7 Oct 2026

Native Parallel Reasoner: Reasoning in Parallelism via Self-Distilled Reinforcement Learning

arXiv 2512.07461 ↗First posted 8 Dec 2025cs.CLcited by 12

On the map: LLM reasoning → Test-time compute and efficient thinking

Abstract (arXiv)

We introduce Native Parallel Reasoner (NPR), a teacher-free framework that enables Large Language Models (LLMs) to self-evolve genuine parallel reasoning capabilities. NPR transforms the model from sequential emulation to native parallel cognition through three key innovations: 1) a self-distilled progressive training paradigm that transitions from ``cold-start'' format discovery to strict topological constraints without external supervision; 2) a novel Parallel-Aware Policy Optimization (PAPO) algorithm that optimizes branching policies directly within the execution graph, allowing the model to learn adaptive decomposition via trial and error; and 3) a robust NPR Engine that refactors memory management and flow control of SGLang to enable stable, large-scale parallel RL training. Across eight reasoning benchmarks, NPR trained on Qwen3-4B achieves performance gains of up to 24.5% and inference speedups up to 4.6x. Unlike prior baselines that often fall back to autoregressive decoding, NPR demonstrates 100% genuine parallel execution, establishing a new standard for self-evolving, efficient, and scalable agentic reasoning.

Key facts. Cited by 12 papers in the map (data through 7 Oct 2026). Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning. Builds on 42 of its 51 references that are in the map. First posted on arXiv 8 Dec 2025.

Builds on

42 of 51 references.

Closest ancestors · 8

Same direction, newest first

  1. 2025A Survey on Parallel Reasoning14 Oct
  2. 2025Training Large Language Models To Reason In Parallel With Global Forking Tokens1 Oct
  3. 2025Parallel-R1: Towards Parallel Thinking via Reinforcement Learning9 Sep
  4. 2025ParaThinker: Native Parallel Thinking as a New Paradigm to Scale LLM Test-time Compute30 Aug
  5. 2025Seek in the Dark: Reasoning via Test-Time Instance-Level Policy Gradient in Latent Space19 May
  6. 2025Learning Adaptive Parallel Reasoning with Language Models21 Apr
  7. 2025s1: Simple test-time scaling31 Jan
  8. 2023Tree of Thoughts: Deliberate Problem Solving with Large Language Models17 May

Other references · 34

  1. 2017Proximal Policy Optimization Algorithmscited by 14,441
  2. 2021Training Verifiers to Solve Math Word Problemscited by 9,743
  3. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 8,851
  4. 2025Qwen3 Technical Reportcited by 8,195
  5. 2025DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learningcited by 6,146
  6. 2023Efficient Memory Management for Large Language Model Serving with PagedAttentioncited by 5,767
  7. 2022Self-Consistency Improves Chain of Thought Reasoning in Language Modelscited by 5,601
  8. 2021Measuring Mathematical Problem Solving With the MATH Datasetcited by 5,203
  9. 2024Qwen2.5 Technical Reportcited by 4,713
  10. 2025Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilitiescited by 3,655
  11. 2023Let's Verify Step by Stepcited by 3,598
  12. 2025DAPO: An Open-Source LLM Reinforcement Learning System at Scalecited by 2,678
22 more
  1. 2022Solving Quantitative Reasoning Problems with Language Modelscited by 1,359
  2. 2023SGLang: Efficient Execution of Structured Language Model Programscited by 1,107
  3. 2024OlympiadBench: A Challenging Benchmark for Promoting AGI with Olympiad-Level Bilingual Multimodal Scientific Problemscited by 1,099
  4. 2024SimPO: Simple Preference Optimization with a Reference-Free Rewardcited by 951
  5. 2023Graph of Thoughts: Solving Elaborate Problems with Large Language Modelscited by 873
  6. 2025Group Sequence Policy Optimizationcited by 788
  7. 2025Open-Reasoner-Zero: An Open Source Approach to Scaling Up Reinforcement Learning on the Base Modelcited by 450
  8. 2025The Lessons of Developing Process Reward Models in Mathematical Reasoningcited by 402
  9. 2025MathArena: Evaluating LLMs on Uncontaminated Math Competitionscited by 275
  10. 2025Logic-RL: Unleashing LLM Reasoning with Rule-Based Reinforcement Learningcited by 260
  11. 2024Monte Carlo Tree Search Boosts Reasoning via Iterative Preference Learningcited by 238
  12. 2025ZebraLogic: On the Scaling Limits of LLMs for Logical Reasoningcited by 186
  13. 2025A Survey of Reinforcement Learning for Large Reasoning Modelscited by 175
  14. 2024Teaching Large Language Models to Reason with Reinforcement Learningcited by 158
  15. 2025Process Reward Models That Thinkcited by 118
  16. 2025Part I: Tricks or Traps? A Deep Dive into RL for LLM Reasoningcited by 53
  17. 2025Thinking vs. Doing: Agents that Reason by Scaling Test-Time Interactioncited by 34
  18. 2025Multiverse: Your Language Models Secretly Decide How to Parallelize and Merge Generationcited by 30
  19. 2025ParallelSearch: Train your LLMs to Decompose Query and Search Sub-queries in Parallel with Reinforcement Learningcited by 24
  20. 2024An Efficient Recipe for Long Context Extension via Middle-Focused Positional Encodingcited by 16
  21. 2025Understanding and Leveraging the Expert Specialization of Context Faithfulness in Mixture-of-Experts LLMscited by 12
  22. 2025RuleReasoner: Reinforced Rule-based Reasoning via Domain-aware Dynamic Samplingcited by 6

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 12.

12 papers.

  1. 2026SCRIBE: Structured Mid-Level Supervision for Tool-Using Language Modelscited by 20
  2. 2026Multi-Stream LLMs: Unblocking Language Models with Parallel Streams of Thoughts, Inputs and Outputscited by 2
  3. 2026On the Overscaling Curse of Parallel Thinking: System Efficacy Contradicts Sample Efficiencycited by 1
  4. 2026Divide-and-Conquer CoT: RL for Reducing Latency via Parallel Reasoningcited by 1
  5. 2026It Takes 8 Tokens: Weak-to-Strong Off-Policy RL via Auxiliary Branchescited by 1
  6. 2026Visual Para-Thinker: Divide-and-Conquer Reasoning for Visual Comprehensioncited by 0
  7. 2026LACE: Lattice Attention for Cross-thread Explorationcited by 0
  8. 2026Learning with a Single Rollout via Monte Carlo Pass@k Criticcited by 0
  9. 2026Fishing Out Free Riders: Shapley-Based Reward Attribution for Parallel Reasoning via Reinforcement Learningcited by 0
  10. 2026BLADE: Boundary-Expanded and Layer-Adaptive Dynamic Exit for Efficient LLM Reasoningcited by 0
  11. 2026Second Thought: Reasoning in Parallel as LLM Agents Act and Observecited by 0
  12. 2026Diversity Combining for Multi-Path LLM Reasoningcited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025ThreadWeaver: Adaptive Threading for Efficient Parallel Reasoning in Language Models
  2. 2025From Efficiency to Adaptivity: A Deeper Look at Adaptive Reasoning in Large Language Models
  3. 2025Motif-2-12.7B-Reasoning: A Practitioner's Guide to RL Training Recipes
  4. 2025RAG-R1: Incentivizing the Search and Reasoning Capabilities of LLMs through Multi-query Parallelism
  5. 2025Adaptive Termination for Multi-round Parallel Reasoning: An Universal Semantic Entropy-Guided Framework
  6. 2026PaCoRe: Learning to Scale Test-Time Compute with Parallel Coordinated Reasoning
  7. 2026ParaTempo: Efficient Parallel Reasoning via Temporal Confidence
  8. 2025Tailored Primitive Initialization is the Secret Key to Reinforcement Learning