Back to AI News
arxiv

Folio: Recoverable Multi-Resolution Context Folding for Long-Horizon Language Agents

Folio is a context manager for long-horizon language agents that records linked turn notes, runtime event manifests, and source records, then folds older history into coarser views. Its fold exposes handles into the manifest and source records so recovery tools can read bounded pages, compute over full files, and keep findings in a persistent dossier. On 75 LOCA-bench tasks with DeepSeek-V4-Pro behind every system, Folio solved 66 tasks versus 54 for the same runtime with model-written compaction and 49-59 for Claude Code, Codex, and SoL-Pi. The paper reports about half their working time, 2.4 times lower cost than that runtime, and a median fold time of 0.6 seconds.

AI Brief

It offers a concrete alternative to model-written context compaction for long-running agents, with benchmark comparisons and reported cost and latency reductions.

arXiv:2611.16627v1 Announce Type: new Abstract: Long-horizon language agents accumulate tool results and replies until the history no longer fits the context window. Current agents shrink this history by having a model summarize it, so each compaction rereads and rewrites the history it replaces, and cache-aware triggers decide when to compact while taking the size of the result as given. We present Folio, a context manager that records, folds, and recovers history through one set of handles. While the agent runs, Folio records three linked levels: a short turn note the model writes as part of its ordinary reply (L1), an event manifest of every call written by the runtime (L2), and the captured source records (L3). Because these views already exist, the runtime folds the context programmatically, selecting coarser views for older turns. The fold prints the handles that index L2 and L3, so recovery tools that read bounded pages, compute over full text in files, and keep findings in a persistent dossier reach the evidence the agent needs with little added input. The fold trigger prices an explicit landing point that includes the retained budget. On 75 LOCA-bench tasks with DeepSeek-V4-Pro behind every system, Folio solves 66 tasks against 54 for the same runtime with model-written compaction and 49-59 for Claude Code, Codex, and SoL-Pi, in about half their working time, at 2.4 times lower cost than that runtime; each fold takes a median of 0.6 seconds.
Be the first to rate this article.

Comments & Questions (0)

Captcha
Please be respectful — let's keep the conversation friendly.

No comments yet

Be the first to comment!