# LLM-Agents-Papers

> LLM-Agents-Papers is a free AI tool for research. Curated, regularly updated reading list of research papers on LLM-based agents. It is available on Web.

- **Kind:** Tool
- **Pricing:** Free
- **Platforms:** Web
- **Repository:** https://github.com/agi-edgerunners/llm-agents-papers
- **GitHub stars:** 2,356
- **Categories:** [research](https://ai-powered.com/categories/research)
- **Listing updated:** 2026-11-21

## About

This repository is a categorized bibliography of academic papers about agents built on large language models. Each entry gives a date, title, a link to the arXiv paper and, where one exists, a link to the authors' code. Entries are sorted by date within each topic, with the newest on top, and the README notes a last-updated date of July 2025.

The papers are grouped into sections: surveys, techniques for enhancement (planning, memory, feedback and reflection, RAG, search), interaction (role playing, conversation, games, tool usage, simulation), application areas such as math, chemistry, medicine, finance and software engineering, automation, training (fine-tuning, RL, DPO), scaling, stability (safety, bias, hallucination) and infrastructure such as benchmarks, environments and datasets.

The README also points to other well-known agent paper lists for wider coverage. It is a reading resource rather than runnable agent software, aimed at researchers and engineers who want to follow the agent literature.

## Features

- **Topic-based organization:** Papers are split into sections such as planning, memory, RAG, tool usage and multi-agent systems so readers can jump to a subject.
- **Survey collection:** A dedicated section gathers survey papers on LLM agents, covering GUI agents, safety, evaluation, medicine, finance and software engineering.
- **Direct paper and code links:** Each entry links to the arXiv paper and to a code repository when one is available.
- **Application-domain coverage:** Dedicated sections track agent research in math, chemistry, biology, physics, geography, art, medicine, finance and research.
- **Infrastructure resources:** Lists benchmarks, evaluation methods, environments, platforms and datasets for agent work.
- **Pointers to related lists:** Recommends other paper collections on LLM agents, self-correction and game agents.

## Good for

- Finding recent surveys to get oriented on LLM agent research
- Searching for papers on agent planning, memory or reflection techniques
- Locating benchmarks and datasets for evaluating agents
- Tracking multi-agent system research for a literature review
- Discovering domain-specific agent papers in medicine, finance or software engineering

## FAQ

### Is this runnable agent software?

No. It is a curated list of research papers about LLM-based agents, not an agent you run.

### How are the papers organized?

By topic, including surveys, planning, memory, feedback and reflection, RAG, search, interaction, applications, training, scaling, stability and infrastructure, each sorted by date.

### Do entries link to source code?

Entries link to the arXiv paper, and also to code when a repository is available.

### How current is the list?

The README states a last-updated date of 2025/7/12.

### Is LLM-Agents-Papers free?

Yes. LLM-Agents-Papers is free to use.

### What are the best alternatives to LLM-Agents-Papers?

The closest alternatives to LLM-Agents-Papers are awesome-trading-agents, DecryptPrompt, ai-agents-from-zero, agents-course and awesome-harness-engineering.

## Where this comes from

- Description: written from the project's README.

## Alternatives

- [awesome-trading-agents](https://ai-powered.com/tools/awesome-trading-agents): awesome-trading-agents is a free AI tool for finance and payments.
- [DecryptPrompt](https://ai-powered.com/tools/decryptprompt): DecryptPrompt is a free AI tool for research.
- [ai-agents-from-zero](https://ai-powered.com/tools/ai-agents-from-zero): ai-agents-from-zero is an open-source AI tool for ai models and llm tooling.
- [agents-course](https://ai-powered.com/tools/agents-course): agents-course is a free AI tool for ai models and llm tooling.
- [awesome-harness-engineering](https://ai-powered.com/tools/awesome-harness-engineering): awesome-harness-engineering is a free AI tool for ai models and llm tooling.
- [awesome-agentic-ai-zh](https://ai-powered.com/tools/awesome-agentic-ai-zh): awesome-agentic-ai-zh is a free AI tool for research.
- [Elicit](https://ai-powered.com/tools/elicit): Elicit is a freemium AI tool for research.
- [rllm](https://ai-powered.com/tools/rllm): rllm is an open-source AI tool for ai models and llm tooling.

---
Source: https://ai-powered.com/tools/llm-agents-papers
