Benchmarking Chat Assistants on Long-Term Interactive Memory (ICLR 2025)

A Multi Agent Memory MCP That Connect Agents Across Systems and Machines

Memory system for AI agents. #1 on LongMemEval — 96.2% (481/500). Beats every published system including Chronos, Mastra, Supermemory, and Emergence. Built solo in 16 days for $1,000.

Memory examples for the Mastra Memory workshop on Jul 24, 2025

No description available

Atlas Memory — self-hosted long-term memory for AI agents. LongMemEval (90.18%)

No description available

High-accuracy long-term memory for AI agents. 86.6% on LongMemEval with 10-step Hyper Search RAG pipeline.

Official Python SDK for RecallrAI – a revolutionary contextual memory system that enables AI assistants to form meaningful connections between conversations, just like human memory.

Visual inspector for LongMemEval dataset

Benchmark suite for conversational memory systems (LongMemEval, ConvoMem)

No description available

LENS - AI Memory Benchmark - Memory as Experience, Not Facts

Local-first AI memory that scores 100% on user fact recall. Open-source memory layer for LLM agents with hybrid search, middle-out compression, and local LLM support. Beats Mem0 (49%) and Zep (71%) on LongMemEval. Your data never leaves your machine.

No description available

No description available

No description available

Neutrally's LongMemEval-S hypothesis file and reproduction instructions — 89.4% (447/500)

在longmemeval评估集上评测

This project provides a complete adaptation of MemoryOS for evaluating long-term memory capabilities on the LongMemEval benchmark, with all core MemoryOS functionality preserved and optimized for evaluation scenarios.

An evaluation adapter for benchmarking the A-mem memory system on LongMemEval, supporting retrieval metrics (Recall@k, NDCG@k) and QA evaluation with multiple LLM backends (OpenAI, SGLang, Ollama).

NOUSai LongMemEval benchmark evidence: 73% with Ollama 3B (local inference)

Implementación de RAG sobre Gemma3:4b. Testeado con LongMemEval

Next-generation agent memory system with 70.4% QA accuracy on LongMemEval

OMEGA persistent memory plugin for OpenClaw — graph-based, local-first, #1 on LongMemEval

Standardized benchmark framework for AI memory systems. Test Mem0, Graphiti, Letta, and more against LongMemEval, LoCoMo, HaluMem.

Open source memory benchmark + RAG system. 82.8% on LongMemEval. Ships as MCP server for Claude Code.

Scalable conversational memory via recursive sub-agent delegation — 46% EM vs 5% truncation on LongMemEval-S, zero training

What works for LLM long-term memory - tested on 2,040 questions across LoCoMo and LongMemEval. CoT prompting > fancy encoding.

A lightweight web app for humans to explore LongMemEval’s long‑context questions and see how the dataset is structured.