MEMENTO is a long‑context agent built on Qwen3‑8B that learns to manage its fixed‑size context window through compaction, replacing its history with self‑generated notes. The model is first fine‑tuned on synthetic trajectories with hindsight notes and then reinforced with a reward combining answer correctness, note quality, and confidence calibration. At test time, MEMENTO can also fold context and backtrack based on confidence, achieving a 67.8% accuracy on book‑level QA (up to 2.5 M tokens) while never holding more than 20 K tokens, outperforming GPT‑5.5 and GLM‑5.3 agents at significantly lower cost.
arXiv:2607. 25066v1 Announce Type: new Abstract: Long-horizon LLM agents accumulate reasoning traces, actions, and tool observations that can eventually exceed a model's fixed context window.
ContextPilot is a proactive context‑management framework designed to improve long‑horizon agentic reasoning with large language models. It expands the toolset to include planning, long‑term memory, and soft context offloading, and introduces a reinforcement‑learning strategy that focuses on critical editing decisions and assigns action‑level advantages. Experiments on long‑context QA and deep search tasks demonstrate that ContextPilot achieves stronger performance with a more compact working context, outperforming existing baselines across various base models and benchmarks.
arXiv:2609.36435v1 Announce Type: new Abstract: An assistant that serves the same user over a long horizon has to answer from what that user has revealed: which preferences still hold, which were rev...
Large Language Model (LLM) agents increasingly rely on external memory systems to accumulate experience across tasks. Yet nearly all existing approaches, from graph-structured memories to reflective insight stores, access memory through fixed, hand-designed heuristics.
arXiv:2607. 13591v1 Announce Type: cross Abstract: Large Language Model (LLM) agents increasingly rely on external memory systems to accumulate experience across tasks.
The paper introduces the Unified Memory Agent (UMA), a system that builds a query‑agnostic external memory from a data stream and reuses it across multiple question‑answering sessions. UMA employs a single policy to manage a structured Memory Bank via CRUD operations and uses Task‑Stratified GRPO to supervise memory maintenance based on QA trajectory rewards. The authors also present Ledger‑QA, a benchmark for long‑horizon state tracking, and demonstrate that UMA outperforms other methods on test‑time learning and accurate‑retrieval tasks, with UMA‑Specialist further improving performance after task adaptation.
arXiv:2502.01562v3 Announce Type: replace Abstract: As the general capabilities of artificial intelligence (AI) agents continue to evolve, their ability to learn to master multiple complex tasks thro...
arXiv:2605. 21850v2 Announce Type: replace-cross Abstract: Recent development of agents has renewed demand for long-context reasoning capacity of LLMs.
arXiv:2608. 07527v1 Announce Type: cross Abstract: Long-document understanding requires models to find and combine evidence across many pages, layouts, tables, figures, and charts.
The paper introduces CTXAgent, a self‑controlled context management system for large language model agents that compacts conversation history when a token threshold is reached. It evaluates seven frontier models, comparing their ability to manage context with and without dedicated tools, and benchmarks performance across coding, scientific, and software engineering tasks. Results show that well‑managed models keep live context below 100k tokens, maintain high benchmark scores, and achieve significant cost reductions, with Claude Opus 5.5 outperforming its explicit configuration by 34% lower cost.
The paper introduces HiMA-Ecom, a hierarchical multi‑agent e‑commerce benchmark with 22.8K instances, where 17.7K include memory entries. It explores training, evaluation, and management of accumulated context in long‑horizon agentic systems, and proposes HiMA‑R1, a policy optimization method that controls trajectory context budgets and uses reward‑driven memory to accelerate convergence. Experiments show that agents built on 3B and 7B open‑source models outperform DeepSeek‑V3 by an average of 6.0 points, while reward‑driven memory cuts convergence steps by about 55%.
Long-horizon tool agents are bottlenecked by how their context grows toward the limits of the context window. Recent systems make context management agent- or system-controlled, but they either learn a compression policy that discards evidence or manage context in a layer the agent never sees.