🤗 HuggingFace ｜ ModelScope | 💬 WeChat(微信) | 📰 Blog | 📑 Paper

👏 Welcome to try Tongyi DeepResearch via our Modelscope online demo or 🤗 Huggingface online demo or bailian service!

[!NOTE] This demo is for quick exploration only. Response times may vary or fail intermittently due to model latency and tool QPS limits. For a stable experience we recommend local deployment; for a production-ready service, visit bailian and follow the guided setup.

Introduction

We present Tongyi DeepResearch, an agentic large language model featuring 30.5 billion total parameters, with only 3.3 billion activated per token. Developed by Tongyi Lab, the model is specifically designed for long-horizon, deep information-seeking tasks. Tongyi DeepResearch demonstrates state-of-the-art performance across a range of agentic search benchmarks, including Humanity's Last Exam, BrowseComp, BrowseComp-ZH, WebWalkerQA,xbench-DeepSearch, FRAMES and SimpleQA.

Tongyi DeepResearch builds upon our previous work on the WebAgent project.

More details can be found in our 📰 Tech Blog.

Features

• ⚙️ Fully automated synthetic data generation pipeline: We design a highly scalable data synthesis pipeline, which is fully automatic and empowers agentic pre-training, supervised fine-tuning, and reinforcement learning.
• 🔄 Large-scale continual pre-training on agentic data: Leveraging diverse, high-quality agentic interaction data to extend model capabilities, maintain freshness, and strengthen reasoning performance.
• 🔁 End-to-end reinforcement learning: We employ a strictly on-policy RL approach based on a customized Group Relative Policy Optimization framework, with token-level policy gradients, leave-one-out advantage estimation, and selective filtering of negative samples to stabilize training in a non‑stationary environment.
• 🤖 Agent Inference Paradigm Compatibility: At inference, Tongyi DeepResearch is compatible with two inference paradigms: ReAct, for rigorously evaluating the model's core intrinsic abilities, and an IterResearch-based 'Heavy' mode, which uses a test-time scaling strategy to unlock the model's maximum perform