MemStrata: 95% and 90.91% Source-Aware Accuracy on LongMemEval-500 and LoCoMo-1540 with a Local Qwen 3.8 27B Q4 K M Reader

arXiv:2610.05343 · cs.CL, cs.IR · Submitted 2026-10-04 · Read on arXiv

cs.CL, cs.IR

Submitted: 2026-10-04

Updated: 2026-10-04

Comments: 29 pages, 23 tables, 1 figure. Ancillary files contain per-question grades and reproducible analyses, plus explicitly labelled exports from audited follow-up reports

Code: https://github.com/yadu9989/memstrata-lme500-harne

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Accuracy — 1× in this paper · explained in 1 episode

Sources

• Mem0: Building Production-Ready AI Agents with Scalable Long-Term Memory
• Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks
• Nomic Embed: Training a Reproducible Long Context Text Embedder
• MemGPT: Towards LLMs as Operating Systems
• Zep: A Temporal Knowledge Graph Architecture for Agent Memory
• Beyond a Million Tokens: Benchmarking and Enhancing Long-Term Memory in LLMs
• LongMemEval: Benchmarking Chat Assistants on Long-Term Interactive Memory
• Temporal Validity in Retrieval Memory: Eliminating Stale-Fact Errors for AI Agents over Evolving Knowledge

Related papers

• Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
• Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
• Subliminal Steering: Stronger Encoding of Hidden Signals
• MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
• The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
• Untangling the Mechanisms of Misleading Context in Medical Question Answering