AdaReasoner: Dynamic Tool Orchestration for Iterative Visual Reasoning

• URL: http://arxiv.org/abs/2601.18631v2
• Date: Wed, 28 Jan 2026 05:44:53 GMT
• Title: AdaReasoner: Dynamic Tool Orchestration for Iterative Visual Reasoning
• Authors: Mingyang Song, Haoyu Sun, Jiawei Gu, Linjie Li, Luxin Xu, Ranjay Krishna, Yu Cheng,
• Abstract summary: We introduce textbfAdaReasoner, a family of multimodal models that learn tool use as a general reasoning skill rather than as tool-specific or explicitly supervised behavior.<n>AdaReasoner is enabled by (i) a scalable data curation pipeline exposing models to long-horizon, multi-step tool interactions; (ii) Tool-GRPO, a reinforcement learning algorithm that prioritizes tool selection and sequencing based on end-task success; and (iii) an adaptive learning mechanism that dynamically regulates tool usage.
• Score: 66.24374176797075
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: When humans face problems beyond their immediate capabilities, they rely on tools, providing a promising paradigm for improving visual reasoning in multimodal large language models (MLLMs). Effective reasoning, therefore, hinges on knowing which tools to use, when to invoke them, and how to compose them over multiple steps, even when faced with new tools or new tasks. We introduce \textbf{AdaReasoner}, a family of multimodal models that learn tool use as a general reasoning skill rather than as tool-specific or explicitly supervised behavior. AdaReasoner is enabled by (i) a scalable data curation pipeline exposing models to long-horizon, multi-step tool interactions; (ii) Tool-GRPO, a reinforcement learning algorithm that optimizes tool selection and sequencing based on end-task success; and (iii) an adaptive learning mechanism that dynamically regulates tool usage. Together, these components allow models to infer tool utility from task context and intermediate outcomes, enabling coordination of multiple tools and generalization to unseen tools. Empirically, AdaReasoner exhibits strong tool-adaptive and generalization behaviors: it autonomously adopts beneficial tools, suppresses irrelevant ones, and adjusts tool usage frequency based on task demands, despite never being explicitly trained to do so. These capabilities translate into state-of-the-art performance across challenging benchmarks, improving the 7B base model by +24.9\% on average and surpassing strong proprietary systems such as GPT-5 on multiple tasks, including VSP and Jigsaw.

Related papers

• Learning to Rewrite Tool Descriptions for Reliable LLM-Agent Tool Use [21.666294374943178]
  We propose a curriculum learning framework that transfers supervision from trace-rich settings to trace-free deployment.<n> Experiments show consistent gains on unseen tools, strong cross-domain generalization, and robustness as the number of candidate tools scales to over 100.
  arXiv  Detail & Related papers  (2026-02-23T23:50:24Z)
• ToolTok: Tool Tokenization for Efficient and Generalizable GUI Agents [16.06309106596998]
  ToolTok is a novel paradigm of multi-step pathfinding for GUI agents.<n>We devise tools aligned with human interaction habits and represent each tool using learnable token embeddings.<n>We construct an easy-to-hard curriculum consisting of three tasks: token definition question-answering, pure text-guided tool selection, and simplified visual pathfinding.
  arXiv  Detail & Related papers  (2026-01-30T08:38:05Z)
• Acting Less is Reasoning More! Teaching Model to Act Efficiently [87.28134636548705]
  Tool-integrated reasoning augments large language models with the ability to invoke external tools to solve tasks.<n>Current approaches typically optimize only for final correctness without considering the efficiency or necessity of external tool use.<n>We propose a framework that encourages models to produce accurate answers with minimal tool calls.<n>Our approach reduces tool calls by up to 68.3% and improves tool productivity by up to 215.4%, while maintaining comparable answer accuracy.
  arXiv  Detail & Related papers  (2025-04-21T05:40:05Z)
• Adaptive Tool Use in Large Language Models with Meta-Cognition Trigger [49.81945268343162]
  We propose MeCo, an adaptive decision-making strategy for external tool use.<n>MeCo quantifies metacognitive scores by capturing high-level cognitive signals in the representation space.<n>MeCo is fine-tuning-free and incurs minimal cost.
  arXiv  Detail & Related papers  (2025-02-18T15:45:01Z)
• ToolGen: Unified Tool Retrieval and Calling via Generation [34.34787641393914]
  We introduce ToolGen, a paradigm shift that integrates tool knowledge directly into the large language models' parameters.<n>We show that ToolGen achieves superior results in both tool retrieval and autonomous task completion.<n>ToolGen paves the way for more versatile, efficient, and autonomous AI systems.
  arXiv  Detail & Related papers  (2024-10-04T13:52:32Z)
• MetaTool: Facilitating Large Language Models to Master Tools with Meta-task Augmentation [25.360660222418183]
  We present MetaTool, a novel tool learning methodology designed to generalize across any reusable toolset. By incorporating meta-task data into task-oriented training, our method significantly enhances the performance of open-source Large Language Models.
  arXiv  Detail & Related papers  (2024-07-15T10:15:41Z)
• Towards Completeness-Oriented Tool Retrieval for Large Language Models [60.733557487886635]
  Real-world systems often incorporate a wide array of tools, making it impractical to input all tools into Large Language Models. Existing tool retrieval methods primarily focus on semantic matching between user queries and tool descriptions. We propose a novel modelagnostic COllaborative Learning-based Tool Retrieval approach, COLT, which captures not only the semantic similarities between user queries and tool descriptions but also takes into account the collaborative information of tools.
  arXiv  Detail & Related papers  (2024-05-25T06:41:23Z)
• Large Language Models as Tool Makers [85.00361145117293]
  We introduce a closed-loop framework, referred to as LLMs A s Tool Makers (LATM), where LLMs create their own reusable tools for problem-solving. Our approach consists of two phases: 1) tool making: an LLM acts as the tool maker that crafts tools for a set of tasks. 2) tool using: another LLM acts as the tool user, which applies the tool built by the tool maker for problem-solving.
  arXiv  Detail & Related papers  (2023-05-26T17:50:11Z)
• Making Language Models Better Tool Learners with Execution Feedback [36.30542737293863]
  Tools serve as pivotal interfaces that enable humans to understand and reshape the environment. Existing tool learning methodologies induce large language models to utilize tools indiscriminately. We propose Tool leaRning wIth exeCution fEedback (TRICE), a two-stage end-to-end framework that enables the model to continually learn through feedback derived from tool execution.
  arXiv  Detail & Related papers  (2023-05-22T14:37:05Z)
• ART: Automatic multi-step reasoning and tool-use for large language models [105.57550426609396]
  Large language models (LLMs) can perform complex reasoning in few- and zero-shot settings. Each reasoning step can rely on external tools to support computation beyond the core LLM capabilities. We introduce Automatic Reasoning and Tool-use (ART), a framework that uses frozen LLMs to automatically generate intermediate reasoning steps as a program.
  arXiv  Detail & Related papers  (2023-03-16T01:04:45Z)

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.