TInR: Exploring Tool-Internalized Reasoning in Large Language Models

Qiancheng Xu; Yongqi Li; Fan Liu; Hongru Wang; Min Yang; Wenjie Li

cs.CLApr 12, 2026

TInR: Exploring Tool-Internalized Reasoning in Large Language Models

Authors: Qiancheng Xu, Yongqi Li, Fan Liu, Hongru Wang, Min Yang, Wenjie Li

Organizations: The Hong Kong Polytechnic University · Southeast University · University of Edinburgh · Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences

Abstract

Tool-Integrated Reasoning (TIR) has emerged as a promising direction by extending Large Language Models' (LLMs) capabilities with external tools during reasoning. Existing TIR methods typically rely on external tool documentation during reasoning. However, this leads to tool mastery difficulty, tool size constraints, and inference inefficiency. To mitigate these issues, we explore Tool-Internalized Reasoning (TInR), aiming at facilitating reasoning with tool knowledge internalized into LLMs. Achieving this goal presents notable requirements, including tool internalization and tool-reasoning coordination. To address them, we propose TInR-U, a tool-internalized reasoning framework for unified reasoning and tool usage. TInR-U is trained through a three-phase pipeline: 1) tool internalization with a bidirectional knowledge alignment strategy; 2) supervised fine-tuning warm-up using high-quality reasoning annotations, and 3) reinforcement learning with TInR-specific rewards. We comprehensively evaluate our method across in-domain and out-of-domain settings. Experiment results show that TInR-U achieves superior performance in both settings, highlighting its effectiveness and efficiency.

Figures & tables

Figure 1: Comparison between (a) Tool-Integrated Reasoning (TIR) and (b) Tool-Internalized Reasoning (TInR). TInR internalizes tool knowledge into LLMs to facilitate reasoning.

Figure 2: Illustration of our proposed TInR-U, with a three-phase training pipeline including tool internalization, TInR SFT warm-up, and TInR RL. TInR-U facilitates tool knowledge internalization and tool usage during reasoning.

Settings	Split	# Instructions	# Tools
In-domain	Train	2552	2467
	Test(Seen)	185	307
	Test(Unseen)	580	831
Out-of-domain		1996	2025

Table 1: Statistics of the experiment datasets conducted from ToolACE Liu et al. (2025a) , xLAM Zhang et al. (2025) , and BFCL Patil et al. (2025) .

Methods	Seen					Unseen
	Tool Identification		Tool Calling			Tool Identification		Tool Calling
	EM	F1	EM	T. Acc	P. Acc	EM	F1	EM	T. Acc	P. Acc
BM25+ToolRL	$48.11$	$54.68$	$54.59$	$65.41$	$61.62$	$42.76$	$51.24$	$46.38$	$59.48$	$51.90$
Ada-embedding+ToolRL	$56.22$	$61.71$	$58.38$	$69.19$	$67.57$	$51.38$	$60.19$	$45.69$	$60.00$	$51.90$
TR-Feedback+ToolRL	$65.41$	$72.61$	$62.70$	$73.51$	$71.35$	$59.31$	$67.80$	$51.38$	$67.41$	$57.76$
ToolRetriever+Hammer2.1-7b	$63.78$	$69.46$	$36.22$	$44.32$	$41.08$	$59.66$	$68.47$	$33.28$	$43.45$	$34.83$

Table 2: In-domain evaluation results of baselines and TInR-U. EM, T.Acc, and P. Acc stand for Exact Match, Tool Accuracy, Parameter Accuracy, respectively. % improve represents the relative improvement achieved by our method over the previously best-performing method.

Methods	Tool Identification		Tool Calling
Methods	EM	F1	EM	T. Acc	P. Acc
BM25+ToolRL	$24.94$	$25.80$	$16.10$	$32.55$	$26.46$
Ada-embedding+ToolRL	$32.32$	$32.38$	$16.04$	$36.82$	$28.92$
TR-Feedback+ToolRL	$30.33$	$30.43$	$16.39$	$35.71$	$29.63$
ToolRetriever+xLAM-7b-r	$30.56$	$30.66$	$10.13$	$27.17$	$21.55$
ToolRetriever+Hammer2.1-7b	$30.56$	$30.66$	$12.06$	$28.57$	$22.72$

Table 3: Out-of-domain evaluation results of baselines and TInR-U.

Methods	Tool Identification		Tool Calling
Methods	EM	F1	EM	T. Acc	P. Acc
TInR-U	78.30	83.44	61.31	77.25	66.27
w/o BKA	$49.67$	$56.81$	$40.39$	$48.63$	$49.15$
w/o RL	$76.47$	$81.46$	$59.61$	$74.90$	$64.18$
w/o recall	$76.21$	$82.05$	$59.74$	$75.29$	$64.58$
w/o memorization	$58.43$	$65.45$	$45.49$	$57.25$	$54.64$

Table 4: Ablation study on key components of TInR-U. BKA stands for bidirectional knowledge alignment.

Figure 3: Comparison of inference efficiency of ToolRL and TInR-U under varying tool set sizes, measured in terms of instructions processed per minute. As the tool size increases, TInR-U demonstrates superior efficiency.

Methods	Tool Identification		Tool Calling
Methods	EM	F1	EM	T. Acc	P. Acc
ToolGen (Qwen)	$76.21$	$82.05$	$59.74$	$75.29$	$64.58$
TInR-U (Qwen)	78.30	83.44	61.31	77.25	66.27
ToolGen (LLaMA)	$70.59$	$76.44$	$51.37$	$67.58$	$57.12$
TInR-U (LLaMA)	75.03	80.32	56.60	73.33	61.57
ToolGen (Mistral)	$70.46$	$77.97$	$51.90$	$69.14$	$57.07$

Table 5: Analysis of TInR-U on different base models.

Methods	Tool Identification		Tool Calling
Methods	EM	F1	EM	T. Acc	P. Acc
Numeric	$42.59$	$49.83$	$16.38$	$41.03$	$30.00$
Hierarchical	$46.72$	$52.24$	$19.14$	$43.14$	$32.07$
Semantic	$50.07$	$56.32$	$20.78$	$48.10$	$33.20$
TInR-U	78.30	83.44	61.31	77.25	66.27

Table 6: Analysis on different internalization methods.

Methods	Tool Identification	Tool Calling
DeepAgent	$62.35$	$55.82$
ToolRetriever+ToolRL	$60.78$	$53.99$
TInR-U	78.30	61.31

Table 7: Comparison with DeepAgent.

Methods	Tool Identification		Tool Calling
Methods	ToolACE	BFCL	ToolACE	BFCL
ToolGen	$73.97$	$31.03$	$58.90$	$22.98$
TInR-U	75.34	34.48	61.64	24.13

Table 8: Multi-turn tool-use results on ToolACE and BFCL datasets.

Methods	Task 1	Task 2
	EM	EM
$\theta_{1}$	$60.55$	$15.53$
$\theta_{2}$	$58.29$	$60.52$

Table 9: Experimental results in continual learning of TInR-U.

Appendix figures & tables5 assets

Supplementary material from the paper’s appendix.

Appendix

Panel of Figure 4 — Figure 4: A case study comparing TInR-U against its variant without bidirectional knowledge alignment.

Methods	Tool Calling
	EM	T. Acc	P. Acc
two-step	$61.31$	$77.25$	$66.27$
w/o recall	$59.74$ (- $2.6\%$ )	$75.29$ (- $2.5\%$ )	$64.58$ (- $2.6\%$ )
single-step	$43.40$	$72.94$	$50.72$
w/o recall	$40.13$ (- $7.5\%$ )	$70.20$ (- $3.8\%$ )	$46.67$ (- $8.0\%$ )

Appendix

Table 10: Ablation of recall in two-step and single-step design of TInR-U on tool calling performance.

Methods	NVIDIA 5880	NVIDIA A6000
ToolRL	[ $2.13,5.10$ ]	[ $2.07,4.86$ ]
TInR-U	$9.29$	$8.97$

Appendix

Table 11: Comparison of inference efficiency of ToolRL and TInR-U with toolsets ranging from 200 to 500 tools across different hardware.

Methods	Tool Identification	Tool Calling
	EM	EM
$\alpha$ =0.5, $\beta$ =1.0	$76.73$	$59.74$
$\alpha$ =2.0, $\beta$ =1.0	$77.51$	$60.52$
$\alpha$ =1.0, $\beta$ =0.5	$77.12$	$58.69$
$\alpha$ =1.0, $\beta$ =2.0	$78.56$	$61.05$
$\alpha$ =1.0, $\beta$ =1.0	$78.30$	$61.31$

Appendix

Table 12: Sensitivity analysis of hyperparameters in TInR-U.

Panel of Figure 5 — Figure 5: The prompt for TInR inference.

LLM Tool Use

Large language model tool use studies how models select, call, and coordinate external tools to extend their reasoning and take actions.

See all 289 papers

PruneTIR: Inference-Time Tool Call Pruning for Effective yet Efficient Tool-Integrated Reasoning
May 11, 2026Luan Zhang, Dandan Song, Zhijing Wu +8LLM Inference Efficiency Tool-Integrated Reasoning
School of Computer Science and Technology, Beijing Institute of Technology, China · 3Independent, China · School of Computer Science and Technology, Huazhong University of Science and Technology, China
Teaching Thinking Models to Reason with Tools: A Full-Pipeline Recipe for Tool-Integrated Reasoning
May 7, 2026Qianjia Cheng, Yuchen Zhang, Zhilin Wang +9Tool-Integrated Reasoning RL for Language Model Reasoning
Zhejiang University · Shanghai AI Laboratory · Peking University +4
DeepTool: Scaling Interleaved Deliberation in Tool-Integrated Reasoning via Process-Supervised Reinforcement Learning
May 28, 2026Yang He, Xiao Ding, Bibo Cai +5Reinforcement Learning Tool-Integrated Reasoning
Research Center for Social Computing and Interactive Robotics, Harbin Institute of Technology, China

May 11, 2026J/K move · Enter open · S save

PruneTIR: Inference-Time Tool Call Pruning for Effective yet Efficient Tool-Integrated Reasoning

Luan Zhang, Dandan Song, Zhijing Wu, Zhengyu Chen+7

School of Computer Science and Technology, Beijing Institute of Technology, China · 3Independent, China · School of Computer Science and Technology, Huazhong University of Science and Technology, China

LLM Inference Efficiency Tool-Integrated Reasoning Inference-Time Optimization LLM Tool Use

Open paper AskPDF arXiv

Tool-integrated reasoning (TIR) enables large language models (LLMs) to enhance their capabilities by interacting with external tools, such as code interpreters (CI). Most recent studies focus on exploring various methods to equip LLMs with the ability to use tools. However, how to further boost the reasoning ability of already tool-capable LLMs at inference time remains underexplored. Improving reasoning at inference time requires no additional training and can help LLMs better leverage tools to solve problems. We observe that, during tool-capable LLM inference, both the number and the proportion of erroneous tool calls are negatively correlated with answer correctness. Moreover, erroneous tool calls are typically resolved successfully within a few subsequent turns. If not, LLMs often struggle to resolve such errors even with many additional turns. Building on the above observations, we propose PruneTIR, a rather effective yet efficient framework that enhances the tool-integrated reasoning at inference time. During LLM inference, PruneTIR prunes trajectories, resamples tool calls, and suspends tool usage through three components: Success-Triggered Pruning, Stuck-Triggered Pruning and Resampling, and Retry-Triggered Tool Suspension. These three components enable PruneTIR to mitigate the negative impact of erroneous tool calls and prevent LLMs from getting stuck in repeated failed resolution attempts, thereby improving overall LLM performance. Extensive experimental results demonstrate the effectiveness of PruneTIR, which significantly improves Pass@1 and efficiency while reducing the working context length for tool-capable LLMs.

Abstract

LLM Tool Use

Explore similar work

PruneTIR: Inference-Time Tool Call Pruning for Effective yet Efficient Tool-Integrated Reasoning

Teaching Thinking Models to Reason with Tools: A Full-Pipeline Recipe for Tool-Integrated Reasoning

DeepTool: Scaling Interleaved Deliberation in Tool-Integrated Reasoning via Process-Supervised Reinforcement Learning