Agent / long-term memory

Most-superseded baselines

Ranked by how many distinct papers critique or beat each method. These are the standard baselines newer work routinely measures against.

1. 1

   Mem0· Mem0

   Mem0: Building Production-Ready AI Agents with Scalable Long-Term Memory

   3 papers critique it · 15 beat it on benchmarks

2. 2

   A-MEM· Mem0

   A-MEM: Agentic Memory for LLM Agents

   5 papers critique it · 12 beat it on benchmarks

3. 3

   Zep· Mem0

   Zep: A Temporal Knowledge Graph Architecture for Agent Memory

   1 papers critique it · 9 beat it on benchmarks

4. 4

   Nemori· Nemori

   Nemori: Self-Organizing Agent Memory Inspired by Cognitive Science

   2 papers critique it · 5 beat it on benchmarks

5. 5

   RAPTOR· RAPTOR

   RAPTOR: Recursive Abstractive Processing for Tree-Organized Retrieval

   3 papers critique it · 4 beat it on benchmarks

6. 6

   MemGPT· RAPTOR

   MemGPT: Towards LLMs as Operating Systems

   4 papers critique it · 3 beat it on benchmarks

7. 7

   HippoRAG 2· RAPTOR

   From RAG to Memory: Non-Parametric Continual Learning for Large Language Models

   1 papers critique it · 5 beat it on benchmarks

8. 8

   MemoryOS· Mem0

   Memory OS of AI Agent

   2 papers critique it · 4 beat it on benchmarks

9. 9

   HippoRAG· RAPTOR

   HippoRAG: Neurobiologically Inspired Long-Term Memory for Large Language Models

   4 papers critique it · 2 beat it on benchmarks

10. 10

    LightMem· Mem0

    LightMem: Lightweight and Efficient Memory-Augmented Generation

    1 papers critique it · 4 beat it on benchmarks

11. 11

    Reflexion· Reflexion

    Reflexion: Language Agents with Verbal Reinforcement Learning

    3 papers critique it · 2 beat it on benchmarks

12. 12

    Memory-R1· Mem0

    Memory-R1: Enhancing Large Language Model Agents to Manage and Utilize Memories via Reinforcement Learning

    4 papers critique it · 1 beat it on benchmarks