MidTool: Mid-training Data Synthesis for Agentic Tool Use

Forum topic · 小凯 · 2026-08-22

Summary

MidTool is an open corpus construction pipeline for mid-training large language models on general agentic tool use, presented in arXiv paper 2608.20314 by Fengqing Jiang, Yite Wang, Boyi Liu, and Zhaoyang Wang. The approach combines large-scale web, PDF, and code data with synthetic supervision derived from real-world tool APIs, MCP skills, and documentation-anchored workflows. MidTool teaches models to recognize tool functionality, ground parameters in context, compose tool-calling workflows, and recover from incomplete information. The authors perform mid-training on Qwen3-4B-Base and Qwen3-8B-Base, followed by supervised fine-tuning and reinforcement learning. Experiments show that the resulting MidTool-Mix consistently improves downstream performance on BFCL, tau2-Bench, and MCP Universe compared with baselines. The results demonstrate that general tool use, like other key LLM capabilities, benefits from dedicated mid-training rather than being left entirely to post-training.

Paper Overview

Research Area: ML Authors: Fengqing Jiang, Yite Wang, Boyi Liu, Zhaoyang Wang Published: 2026-08-22 arXiv: 2608.20314

Abstract

Mid-training is increasingly recognized as a critical stage for shaping the capabilities of large language models. Recent work shows that targeted mid-training can enhance reasoning-intensive abilities and also improve agentic capabilities in software engineering environments. This paper investigates a parallel but less-explored agentic capability: general tool use.

The authors propose MidTool, an open corpus construction pipeline for mid-training on agentic tool use. It combines large-scale web, PDF, and code data with synthetic supervision derived from real-world tool APIs, MCP skills, and documentation-anchored workflows. MidTool is designed to teach models to:

Recognize tool functionality
Ground parameters from context
Compose tool-calling workflows
Recover from incomplete information

The models are mid-trained on Qwen3-4B-Base and Qwen3-8B-Base, then post-trained with SFT and RL.

Results

Compared with baselines, MidTool-Mix consistently improves downstream performance on BFCL, tau2-Bench, and MCP Universe. The results indicate that general tool use, like other important LLM capabilities, benefits from dedicated mid-training rather than being left entirely to post-training.

---

*Auto-collected on 2026-08-22*