{"slug": "stateless-agents-cut-long-horizon-token-usage", "title": "Stateless agents cut long‑horizon token usage", "summary": "A paper on Stateless Language Agents (SLAs) reports that decoupling durable state from the model's active context lets long-horizon agents reach the strongest kernel baseline's final performance while using over 84% fewer tokens. The approach stores candidate solutions and outcomes in an external harness that feeds each stateless worker only the concise evidence it needs, with a coordinating Advisor consuming under 0.6% of tokens. The gains held across software engineering, kernel optimization and algorithm design experiments at budgets up to one billion tokens, though evaluation was limited to three engineered domains.", "body_md": "Current long‑horizon LLM agents burn tokens at a rate that makes billion‑step research infeasible. Their prompts grow as every intermediate result is replayed, leading to duplicated effort and runaway costs. Decoupling durable state from the model’s active context, as proposed by Stateless Language Agents (SLAs), flips this trend on its head.\n\nHistorically, agent designs have embedded the entire research history inside each model invocation, so every call re‑feeds an ever‑larger transcript. Benchmarks were consequently limited to short budgets where token blow‑up never surfaced. This coupling forced researchers to accept a trade‑off between horizon length and compute expense.\n\nStateless Language Agents reach the strongest kernel baseline’s final performance while using over 84 % fewer tokens. By storing candidate solutions and outcomes in an external harness, each agent receives only the concise evidence it needs for its next step. The paper reports that “SLA achieves the best final result on every task and reaches the strongest kernel baseline’s final performance with over 84% fewer tokens” [\\[1\\]](https://arxiv.org/abs/2610.07625). This reduction holds across software engineering, kernel optimization, and algorithm design experiments at budgets up to one billion tokens.\n\nThe coordination component – the Advisor – consumes less than 0.6 % of tokens, keeping overhead negligible. The harness builds a role‑specific context for each stateless worker, while the Advisor merely reads summarized evidence and issues concrete experiment assignments. As quoted, “while the Advisor consumes less than 0.6% of tokens” [\\[1\\]](https://arxiv.org/abs/2610.07625), meaning almost all budget is spent on substantive reasoning rather than bookkeeping.\n\nStateless agents still depend on a harness that must summarize evidence accurately, and the paper’s evaluation is limited to three engineered domains. If the summarizer drops critical nuance, workers may pursue suboptimal directions despite the token savings. Moreover, tasks lacking clear candidate‑outcome structures could resist this externalization, leaving open whether the same gains translate to more unstructured research settings.\n\nFuture research pipelines should externalize durable state and re‑evaluate long‑horizon benchmarks with token budgets in the billions. By treating the harness as the sole keeper of history, practitioners can scale experiments on fixed‑size models without incurring prohibitive prompt costs. This shift will likely render many current “short‑budget” results obsolete and set a new baseline for cost‑effective automated research.", "url": "https://wpnews.pro/news/stateless-agents-cut-long-horizon-token-usage", "canonical_source": "https://dev.to/olaughter/stateless-agents-cut-long-horizon-token-usage-oog", "published_at": "2026-10-24 05:00:00+00:00", "updated_at": "2026-10-24 05:11:33.346088+00:00", "lang": "en", "topics": ["ai-agents", "large-language-models", "ai-research", "ai-infrastructure"], "entities": ["Stateless Language Agents", "Advisor"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/stateless-agents-cut-long-horizon-token-usage", "markdown": "https://wpnews.pro/news/stateless-agents-cut-long-horizon-token-usage.md", "text": "https://wpnews.pro/news/stateless-agents-cut-long-horizon-token-usage.txt", "jsonld": "https://wpnews.pro/news/stateless-agents-cut-long-horizon-token-usage.jsonld"}}