Repolode
Submit
weitianxin

weitianxin/Awesome-Agentic-Reasoning

A curated paper list accompanying a survey on agentic reasoning in LLMs, covering planning, tool use, search, memory, multi-agent systems and applications.

View on GitHubWebsite
Stars
1,427
Forks
115
Open issues
16
Last commit
2 months ago

What is it?

Awesome-Agentic-Reasoning is a curated collection of research papers and resources built around the survey "Agentic Reasoning for Large Language Models." It organizes work on LLM agents that combine reasoning with action, arranged into foundational single-agent abilities, self-evolving reasoning through feedback and memory, and collective reasoning in multi-agent systems.

The list groups papers by theme, such as planning (in-context workflows, tree search, post-training planning), tool-use optimization, agentic search, feedback mechanisms, and agentic memory. It also covers application areas including math and coding agents, scientific discovery, embodied agents, healthcare, and web research agents, plus a benchmarks section.

The repository accepts pull requests and issues for new papers, and it links to the arXiv version of the survey along with a citation entry.

Key features

  • 01Papers grouped into foundational, self-evolving, and collective agentic reasoning layers
  • 02Coverage of planning, tool use, agentic search, and agentic memory
  • 03Dedicated sections on multi-agent roles and collaboration
  • 04Application area lists for coding, science, embodied, medical, and web agents
  • 05Benchmarks section for evaluating agentic reasoning

Use cases

  • Researchers looking for prior work on LLM agent planning or tool use
  • Engineers scanning recent papers on agent memory and self-improvement
  • Students building a reading list on agentic AI

In collections

Open-source AI agent frameworks & apps

Topics

#agent#llm#reasoning#survey#paper-list#multi-agent#tool-use#agentic-ai#awesome-resources#generative-ai

FAQ

What is Awesome-Agentic-Reasoning?
A curated paper list accompanying a survey on agentic reasoning in LLMs, covering planning, tool use, search, memory, multi-agent systems and applications. Awesome-Agentic-Reasoning is a curated collection of research papers and resources built around the survey "Agentic Reasoning for Large Language Models." It organizes work on LLM agents that combine reasoning with action, arranged into foundational single-agent abilities, self-evolving reasoning through feedback and memory, and collective reasoning in multi-agent systems.
Is Awesome-Agentic-Reasoning free to use? What license does it have?
Awesome-Agentic-Reasoning is open source under the MIT license. Check the license terms before using it in commercial products.
Is Awesome-Agentic-Reasoning still maintained?
Updated within 6 months. The last commit to Awesome-Agentic-Reasoning was on Sep 21, 2026.
What can I use Awesome-Agentic-Reasoning for?
Researchers looking for prior work on LLM agent planning or tool use Engineers scanning recent papers on agent memory and self-improvement Students building a reading list on agentic AI
Where can I get Awesome-Agentic-Reasoning?
The source code and installation instructions are on GitHub: https://github.com/weitianxin/Awesome-Agentic-Reasoning

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects