Awesome-Agentic-Reasoning

Why we included this project

Researchers and engineers trying to get a structured view of how large language models move from pure text reasoning into acting on the world will find this curated paper index a good starting point. It is built around the survey "Agentic Reasoning for Large Language Models" and organizes the field into three layers: foundational single-agent abilities like planning, tool use, and search; self-evolving behavior through memory and feedback; and collective multi-agent coordination. Each layer has its own section with papers mapped onto it, so instead of hunting through scattered arXiv listings you get a thematic map with planning, tool-use optimization, agentic search, memory, feedback mechanisms, and multi-agent systems each in their place, plus a dedicated benchmarks area. The applications section covers math, scientific discovery, embodied agents, healthcare, and autonomous web research, which makes it easy to see where a given capability has actually been put to use. The list accepts contributions and keeps expanding as new work appears.