Tongyi DeepResearch

Open-source agentic language model (30.5B parameters, 3.3B active per token) trained for long, multi-step web information seeking.

Agent by Alibaba Tongyi Lab

IndexedApache-2.0★ 20,076

Write a review

Get started

Setup steps are in the repository. Website · Repository

About

Tongyi DeepResearch is a model from Alibaba's Tongyi Lab built for long-horizon, deep information-seeking tasks, with 30.5 billion total parameters and 3.3 billion active per token. It can run under the ReAct paradigm or an IterResearch-based Heavy mode that spends more compute at test time. The project builds on the earlier WebAgent work and offers online demos plus downloadable weights.

Use cases: research, deep-research, web-search

Reviews

Write a review

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Fact sheet

Jobs: researchREADME
Where you use it: WebREADME
Hosting: Hosted or self-hostedREADME
Models: Tongyi-DeepResearch-30B-A3BREADME
Open source: YesREADME

Facts extracted by Claude Sonnet 5.5 from the linked sources on Oct 24, 2026. Wrong fact? Tell us.

Built on

Frameworks named in its dependency files.

0.525score
Qwen-Agent
Alibaba Qwen
Python framework for LLM applications built on Qwen models, with function calling, MCP, a code interpreter and RAG options.
HarnessDFreshness: D★ 17.2K↓ 21K/wk

Signals

Stars: 20,076
Forks: 1,529
Open issues: 81
Last commit: Feb 27, 2026 (8 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License: Apache-2.0
Repository: github.com/Alibaba-NLP/DeepResearch
Website: tongyi-agent.github.io/blog/introducing-tongyi-deep-research/
First seen: Oct 24, 2026
Last checked: Oct 24, 2026
Source: README github-graphql

More from Alibaba Tongyi Lab

Cite this page

Tongyi DeepResearch. Agent Mart. https://agentmart.dev/agents/tongyi-deepresearch (last updated 2026-10-24)