# SLICE-Mem preprint reports lower token use for AI agent memory

_Published Tuesday, October 27, 2026 at 10:05 AM EDT · AI, Science · Latest · Tier 2 — Notable_

Researchers report that SLICE-Mem, a framework for building AI agents' long-term memory, reduced total language-model token consumption by 92.4% on Memora Weekly and 99.5% on LongMemEval-S compared with MemOS in tests using Qwen3-30B.

A compact classifier selects user messages worth retaining, and an extractor converts them into concise memory entries without generative language-model calls. Reconciliation across entries happens later in an offline consolidation stage.

The preprint reports a performance tradeoff: SLICE-Mem scored 46.60 versus MemOS's 46.68 on Memora Weekly's Forgetting-Aware Memory Accuracy measure, and 56.8% versus 59.3% accuracy on LongMemEval-S.

## Sources

- [cs.CL updates on arXiv.org](https://arxiv.org/abs/2610.33015)

---
Canonical: https://techandbusiness.org/newswire/mRuIOHyJjz_0j2YKvNfFa3
Published: 2026-10-27T14:05:53.570Z
Story chronology: 2026-10-27T04:00:00.000Z
Retrieved: 2026-10-27T16:36:18.857Z
Publisher: Tech & Business (techandbusiness.org)