Alibaba-NLP/DeepResearch

Tongyi Deep Research, the Leading Open-source Deep Research Agent

55
/ 100
Established

Implements a 30.5B parameter sparse-activation model (3.3B per token) trained on synthetic agentic data through continual pre-training, supervised fine-tuning, and on-policy reinforcement learning with Group Relative Policy Optimization. Supports dual inference paradigms—ReAct for core capability evaluation and IterResearch-based "Heavy" mode for test-time scaling—with tool integrations for web search (Serper), page reading (Jina), code execution (SandboxFusion), and multi-modal document parsing.

18,422 stars.

No Package No Dependents
Maintenance 10 / 25
Adoption 10 / 25
Maturity 16 / 25
Community 19 / 25

How are scores calculated?

Stars

18,422

Forks

1,419

Language

Python

License

Apache-2.0

Last pushed

Feb 27, 2026

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/llm-tools/Alibaba-NLP/DeepResearch"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.