ASTER: Agentic Scaling with Tool-integrated Extended Reasoning

• URL: http://arxiv.org/abs/2602.01204v1
• Date: Sun, 01 Feb 2026 12:46:02 GMT
• Title: ASTER: Agentic Scaling with Tool-integrated Extended Reasoning
• Authors: Xuqin Zhang, Quan He, Zhenrui Zheng, Zongzhang Zhang, Xu He, Dong Li,
• Abstract summary: Reinforcement learning (RL) has emerged as a dominant paradigm for eliciting long-horizon reasoning in Large Language Models (LLMs)<n>We introduce ASTER (Agentic Scaling with Tool-integrated Extended Reasoning), a framework that circumvents this collapse through a targeted cold-start strategy.<n>We find that a small expert cold-start set of just 4K interaction-dense trajectories yields the strongest downstream performance.
• Score: 27.877412657068806
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Reinforcement learning (RL) has emerged as a dominant paradigm for eliciting long-horizon reasoning in Large Language Models (LLMs). However, scaling Tool-Integrated Reasoning (TIR) via RL remains challenging due to interaction collapse: a pathological state where models fail to sustain multi-turn tool usage, instead degenerating into heavy internal reasoning with only trivial, post-hoc code verification. We systematically study three questions: (i) how cold-start SFT induces an agentic, tool-using behavioral prior, (ii) how the interaction density of cold-start trajectories shapes exploration and downstream RL outcomes, and (iii) how the RL interaction budget affects learning dynamics and generalization under varying inference-time budgets. We then introduce ASTER (Agentic Scaling with Tool-integrated Extended Reasoning), a framework that circumvents this collapse through a targeted cold-start strategy prioritizing interaction-dense trajectories. We find that a small expert cold-start set of just 4K interaction-dense trajectories yields the strongest downstream performance, establishing a robust prior that enables superior exploration during extended RL training. Extensive evaluations demonstrate that ASTER-4B achieves state-of-the-art results on competitive mathematical benchmarks, reaching 90.0% on AIME 2025, surpassing leading frontier open-source models, including DeepSeek-V3.2-Exp.

Related papers

• Internalizing LLM Reasoning via Discovery and Replay of Latent Actions [4.830503861275364]
  Internalization of chain-of-thought processes into hidden states has emerged as a highly efficient paradigm for scaling test-time compute.<n>We propose STIR (Self-Distilled Tools for Internal Reasoning), a framework that reformulates reasoning enhancement as a dynamic latent trajectory control problem.
  arXiv  Detail & Related papers  (2026-02-04T08:44:57Z)
• MiroThinker: Pushing the Performance Boundaries of Open-Source Research Agents via Model, Context, and Interactive Scaling [115.74855199827596]
  MiroThinker is an open-source research agent designed to advance tool-augmented reasoning and information-seeking capabilities.<n>Unlike previous agents that only scale up model size or context length, MiroThinker explores interaction scaling at the model level.
  arXiv  Detail & Related papers  (2025-11-14T18:52:07Z)
• Demystifying Reinforcement Learning in Agentic Reasoning [90.3737088727791]
  We conduct a comprehensive and systematic investigation to demystify reinforcement learning in agentic reasoning.<n>We highlight our key insights: (i) replacing stitched synthetic trajectories with real end-to-end tool-use trajectories yields a far stronger SFT.<n> Exploration-friendly techniques are crucial for agentic RL, such as clip higher, overlong reward shaping, and maintaining adequate policy entropy could improve the training efficiency.
  arXiv  Detail & Related papers  (2025-10-13T17:57:15Z)
• Beneficial Reasoning Behaviors in Agentic Search and Effective Post-training to Obtain Them [23.986035712600657]
  We propose a reasoning-driven pipeline to study effective reasoning behavior patterns in agentic search.<n>We identify four beneficial reasoning behaviors: Information Verification, Authority Evaluation, Adaptive Search, and Error Recovery.<n>We show that behavior priming yields over 35% gains in Llama3.2-3B and Qwen3-1.7B compared to directly training agentic search models with RL.
  arXiv  Detail & Related papers  (2025-10-08T00:20:35Z)
• Learning to Reason as Action Abstractions with Scalable Mid-Training RL [55.24192942739207]
  An effective mid-training phase should identify a compact set of useful actions and enable fast selection.<n>We propose Reasoning as Action Abstractions (RA3), a scalable mid-training algorithm.
  arXiv  Detail & Related papers  (2025-09-30T05:34:20Z)
• HOID-R1: Reinforcement Learning for Open-World Human-Object Interaction Detection Reasoning with Multimodal Large Language Model [13.82578761807402]
  We introduce HOID-R1, the first HOI detection framework that integrates chain-of-thought (CoT) guided fine-tuning with group relative policy optimization.<n>To mitigate hallucinations in the CoT reasoning, we introduce an "MLLM-as-a-judge" mechanism that supervises the CoT outputs.<n>Experiments show that HOID-R1 achieves state-of-the-art performance on HOI detection benchmarks and outperforms existing methods in open-world generalization to novel scenarios.
  arXiv  Detail & Related papers  (2025-08-15T09:28:57Z)
• RL-PLUS: Countering Capability Boundary Collapse of LLMs in Reinforcement Learning with Hybrid-policy Optimization [111.1749164063616]
  We propose RL-PLUS, a novel hybrid-policy optimization approach for Large Language Models (LLMs)<n> RL-PLUS synergizes internal exploitation with external data to achieve stronger reasoning capabilities and surpass the boundaries of base models.<n>We provide both theoretical analysis and extensive experiments to demonstrate the superiority and generalizability of our approach.
  arXiv  Detail & Related papers  (2025-07-31T23:55:29Z)
• Agentic Reinforced Policy Optimization [66.96989268893932]
  Large-scale reinforcement learning with verifiable rewards (RLVR) has demonstrated its effectiveness in harnessing the potential of large language models (LLMs) for single-turn reasoning tasks.<n>Current RL algorithms inadequately balance the models' intrinsic long-horizon reasoning capabilities and their proficiency in multi-turn tool interactions.<n>We propose Agentic Reinforced Policy Optimization (ARPO), a novel agentic RL algorithm tailored for training multi-turn LLM-based agents.
  arXiv  Detail & Related papers  (2025-07-26T07:53:11Z)
• Beyond Accuracy: Dissecting Mathematical Reasoning for LLMs Under Reinforcement Learning [93.00629872970364]
  Reinforcement learning (RL) has become the dominant paradigm for improving the performance of language models on complex reasoning tasks.<n>We introduce SPARKLE, a fine-grained analytic framework to dissect the effects of RL across three key dimensions.<n>We study whether difficult problems -- those yielding no RL signals and mixed-quality reasoning traces -- can still be effectively used for training.
  arXiv  Detail & Related papers  (2025-06-05T07:53:59Z)
• DGRO: Enhancing LLM Reasoning via Exploration-Exploitation Control and Reward Variance Management [18.953750405635393]
  Decoupled Group Reward Optimization (DGRO) is a general RL algorithm for Large Language Models (LLMs) reasoning.<n>We show that DGRO achieves state-of-the-art performance on the Logic dataset with an average accuracy of 96.9%, and demonstrates strong generalization across mathematical benchmarks.
  arXiv  Detail & Related papers  (2025-05-19T10:44:49Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.