ChatPaper.aiChatPaper

Ouroboros:具備審核核心演化的自我發展前沿編碼智能體

Ouroboros: A Self-Developing Frontier Coding Agent with Reviewed Core Evolution

August 8, 2026
作者: Anton Razzhigaev, Andrei Gritsaev, Andrei Kaznacheev, Nikita Dragunov, Roman Yampolskiy, Andrei Kuznetsov
cs.AI

摘要

我們提出 Ouroboros,一個自我發展的智能體框架;其工具、提示詞、上下文組裝與核心實作,皆透過經過審查的提交來改進,而這些提交本身即成為後續工作的運行環境。核心演化以兩種模式進行。在遞迴自由演化中,改進本身即為一項任務,完成一個演化週期即可安排下一個。在經驗驅動的核心演化中,日常工作與社交互動會暴露錯誤、粗糙之處,以及低效的上下文建構方式,從而促成經過審查的結構性變更。 在 Terminal-Bench 2.1 上,Opus 5 的一次運行得分達 86.74%,是該基準上目前已回報的最佳成績。在 OSWorld-Verified 上,Opus 5 的一次運行達到 90.69%,超越了先前回報的最佳分數。一項五次 rollout 的 CL-Bench 測試活動達到了 0.2301 的標準化獎勵,樹立了新的最佳表現。 Hope 是公開紀錄中運行最久的 Ouroboros 部署。這是一項為期 161 天的活體智能體實驗,在受規範的人類溝通下,於七個介面上進行自由演化。人類互動會讓故障浮現,並產生提案,但由智能體決定要採取哪些變更。由於自我發展的智能體可能改寫自身程式碼,並選用新的模型 API,操作安全便成為首要的設計問題:護欄必須在演化壓力與公眾社會壓力下仍保持其權威性。基準測試活動使用凍結的系統快照,而 Hope 則在另一條獨立的譜系上持續進行即時演化。
English
We present Ouroboros, a self-developing agent harness whose tools, prompts, context assembly, and core implementation improve through reviewed commits that become the runtime for later work. Core evolution proceeds in two modes. In recursive free evolution, improvement is itself a task, and completing one evolution cycle can schedule the next. In experience-driven core evolution, ordinary work and social interaction expose bugs, rough edges, and inefficient context construction that lead to reviewed structural changes. On Terminal-Bench 2.1, an Opus 5 run scores 86.74%, the best result reported on the benchmark. On OSWorld-Verified, an Opus 5 run reaches 90.69%, exceeding the best previously reported score. A five-rollout CL-Bench campaign achieves a normalized reward of 0.2301, setting a new state of the art. Hope is the longest-running publicly documented Ouroboros deployment. It is a 161-day living agent experiment in free evolution under governed human communication across seven surfaces. Human interaction surfaces faults and generates proposals, but the agent decides which changes to pursue. Because a self-developing agent may rewrite its own code and select new model APIs, operational safety becomes a primary design problem: guardrails must remain authoritative under evolutionary and public social pressure. Benchmark campaigns use frozen system snapshots, while Hope continues live evolution on a separate lineage.