mlsanityData to 7 Oct 2026

To Call or Not to Call: A Framework to Assess and Optimize LLM Tool Calling

arXiv 2605.00737 ↗First posted 1 May 2026cs.AIcited by 7

On the map: LLM agents and tool use → Tool learning and function calling

Abstract (arXiv)

Agentic AI architectures augment LLMs with external tools, unlocking strong capabilities but potentially incurring substantial costs. Moreover, tool use is not always beneficial: redundant or low-utility calls can even harm task performance. Effective tool use, therefore, hinges on a core LLM decision: whether to call or not call a tool when performing a task. We introduce a principled framework inspired by decision-making theory to understand tool-use decisions along three key factors: necessity, utility, and affordability. Our analysis combines two complementary lenses: a normative perspective that infers true need and utility for optimal tool calls, and a descriptive perspective that infers the model's self-perceived need and utility from their observed behaviors. We evaluate six open models and a proprietary OpenAI model across native and customized harnesses, two tools, and six tasks. Models' perceived need and utility remain misaligned with their true values, particularly under budget constraints. This misalignment produces both costly overuse and performance-degrading calls. To improve the tool decisions, we train lightweight latent estimators of need (LNEs) from model hidden states. LNEs generally predict true need more accurately than model self-reports and improve budgeted tool allocation across model scales and tool types. Code and dataset available at https://github.com/QinyuanWu0710/ToCall_or_NotToCall.

Key facts. Cited by 7 papers in the map (data through 7 Oct 2026). Placed in Tool learning and function calling, one of 16 directions in LLM agents and tool use. Builds on 45 of its 52 references that are in the map. First posted on arXiv 1 May 2026.

Builds on

45 of 52 references.

Closest ancestors · 9

Same direction, newest first

  1. 2025When2Call: When (not) to Call Tools26 Apr
  2. 2024Tool Learning with Large Language Models: A Survey28 May
  3. 2023MetaTool Benchmark for Large Language Models: Deciding Whether to Use Tools and Which to Use4 Oct
  4. 2023ToolLLM: Facilitating Large Language Models to Master 16000+ Real-world APIs31 Jul
  5. 2023GEAR: Augmenting Language Models with Generalizable and Efficient Tool Resolution17 Jul
  6. 2023Gorilla: Large Language Model Connected with Massive APIs24 May
  7. 2023Making Language Models Better Tool Learners with Execution Feedback22 May
  8. 2023API-Bank: A Comprehensive Benchmark for Tool-Augmented LLMs14 Apr
  9. 2023Toolformer: Language Models Can Teach Themselves to Use Tools9 Feb

Other references · 36

  1. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,160
  2. 2022ReAct: Synergizing Reasoning and Acting in Language Modelscited by 8,355
  3. 2023Efficient Memory Management for Large Language Model Serving with PagedAttentioncited by 5,767
  4. 2023Self-RAG: Learning to Retrieve, Generate, and Critique through Self-Reflectioncited by 1,800
  5. 2022Language Models (Mostly) Know What They Knowcited by 1,783
  6. 2025Search-R1: Training LLMs to Reason and Leverage Search Engines with Reinforcement Learningcited by 1,371
  7. 2022When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric Memoriescited by 1,103
  8. 2023CRITIC: Large Language Models Can Self-Correct with Tool-Interactive Critiquingcited by 751
  9. 2023Active Retrieval Augmented Generationcited by 506
  10. 2024Adaptive-RAG: Learning to Adapt Retrieval-Augmented Large Language Models through Question Complexitycited by 437
  11. 2025Search-o1: Agentic Search-Enhanced Large Reasoning Modelscited by 423
  12. 2025ReTool: Reinforcement Learning for Strategic Tool Use in LLMscited by 377
24 more
  1. 2025ToolRL: Reward is All Tool Learning Needscited by 342
  2. 2024Semantic Entropy Probes: Robust and Cheap Hallucination Detection in LLMscited by 246
  3. 2024LLMs Know More Than They Show: On the Intrinsic Representation of LLM Hallucinationscited by 230
  4. 2025ToRL: Scaling Tool-Integrated RLcited by 166
  5. 2025An Illusion of Progress? Assessing the Current State of Web Agentscited by 165
  6. 2024VERISCORE: Evaluating the factuality of verifiable claims in long-form text generationcited by 114
  7. 2025Agentic Reasoning and Tool Integration for LLMs via Reinforcement Learningcited by 107
  8. 2023WebGLM: Towards An Efficient Web-Enhanced Question Answering System with Human Preferencescited by 100
  9. 2024Large Language Models Must Be Taught to Know What They Don't Knowcited by 99
  10. 2024DRAGIN: Dynamic Retrieval Augmented Generation based on the Information Needs of Large Language Modelscited by 63
  11. 2023Knowledge of Knowledge: Exploring Known-Unknowns Uncertainty with Large Language Modelscited by 60
  12. 2023On Early Detection of Hallucinations in Factual Question Answeringcited by 54
  13. 2025Budget-Aware Tool Use Enables Effective Agent Scalingcited by 42
  14. 2024SeaKR: Self-aware Knowledge Retrieval for Adaptive Retrieval Augmented Generationcited by 40
  15. 2024Estimating Knowledge in Large Language Models Without Generating a Single Tokencited by 39
  16. 2025Adaptive Retrieval Without Self-Knowledge? Bringing Uncertainty Back Homecited by 33
  17. 2024Rowen: Adaptive Retrieval-Augmented Generation for Hallucination Mitigation in LLMscited by 27
  18. 2024WildHallucinations: Evaluating Long-form Factuality in LLMs with Real-World Entity Queriescited by 26
  19. 2024When LLMs Go Online: The Emerging Threat of Web-Enabled LLMscited by 22
  20. 2024CATP-LLM: Empowering Large Language Models for Cost-Aware Tool Planningcited by 14
  21. 2025Don't lie to your friends: Learning what you know from collaborative self-playcited by 14
  22. 2025Query-Level Uncertainty in Large Language Modelscited by 12
  23. 2025Characterizing Web Search in The Age of Generative AIcited by 5
  24. 2026Bowling with ChatGPT: On the Evolving User Interactions with Conversational AI Systemscited by 3

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 7.

7 papers.

  1. 2026Exploring Agentic Tool-Calling Decisions via Uncertainty-Aligned Reinforcement Learningcited by 4
  2. 2026Scores Are Not Decisions: Cost-Aware Stopping for Tool Acquisition in LLM Agentscited by 2
  3. 2026Characterizing Web Search by Conversational LLM Agents: From Search Decisions and Strategies to Results and Responsescited by 0
  4. 2026LazyAgent: Demand-Driven Materialization and Physical Optimization of Agentic Programscited by 0
  5. 2026Nudgeability: Reasoning Models Follow Confidence Signals Without Tracking Their Own Competencecited by 0
  6. 2026When Does Correction Become Repair? Mechanistic Auditing of Internal Interventions in Tool-Using LLMscited by 0
  7. 2026What Can Component-Replacement Evidence Establish? A Critical Scoping Review of Local Decisions in LLM Agentscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026LLM Agents Already Know When to Call Tools -- Even Without Reasoning
  2. 2026Model-Adaptive Tool Necessity Reveals the Knowing-Doing Gap in LLM Tool Use
  3. 2026Looking Is Not Picking: An Attention-Segment Account of Tool-Selection Failures in LLM Agents
  4. 2026The Evolution of Tool Use in LLM Agents: From Single-Tool Call to Multi-Tool Orchestration
  5. 2026Tunable Tool-Call Rates in LLM Agents via Representation Steering
  6. 2026AgentJudgeBench: A Multi-Difficulty Benchmark for Evaluating LLM Judges on Agentic Tool-Calling
  7. 2026ToolFailBench: Diagnosing Tool-Use Failures in LLM Agents
  8. 2026Are Tools All We Need? Unveiling the Tool-Use Tax in LLM Agents