主智能体：一种自改进的强化学习模型框架

摘要

语言模型是顺序处理器，但长时程智能体行为需要超越模型权重和活动上下文的外部信息与计算。Prime Agent是一个开源框架，用于长时程评估和编码智能体工作流。持久化IPython REPL遵循递归语言模型抽象，实现程序化上下文处理和测试时计算，而持续框架在轨迹间保留历史、记忆、技能、提示词和子智能体规格。递归子智能体通过智能体间直接通信进行协调，智能体视图允许人类检查和管理由守护进程支持的会话。Prime Agent标准化了执行、恢复、验证和资源核算，同时将策略构建留给模型。这种低摩擦、高表达力的界面防止框架故障演变为模型故障，使测量更接近模型真实的最大潜在能力。Prime Agent将ARC-AGI-3 RHAE Best@1从30%提升至95.5%，并在长上下文编码、GPU内核生成、模拟器构建和自主nanoGPT速通方面达到或超越原生及主流框架。在《异星工厂》中，我们发现精炼机制支持持续的技术进步，而专用子智能体能够实现并行化工作。代码可在https://github.com/PrimeIntellect-ai/prime-agent获取。

English

Language models are sequential processors, but long-horizon agency requires external information and computation beyond model weights and active context. Prime Agent is an open-source harness for long-horizon evaluation and coding-agent workflows. A persistent IPython REPL follows the Recursive Language Model abstraction for programmatic context processing and test-time compute, while Continual Harness preserves histories, memories, skills, prompts, and subagent specifications across trajectories. Recursive subagents coordinate through direct agent-to-agent communication, and the Agents View lets humans inspect and manage daemon-backed sessions. Prime Agent standardizes execution, recovery, verification, and resource accounting while leaving strategy construction to the model. This low-friction, expressive membrane prevents harness failures from becoming model failures and pushes measurement toward the model's true maximal underlying capability. Prime Agent raises ARC-AGI-3 RHAE Best@1 from 30% to 95.5% and matches or exceeds native and popular harnesses across long-context coding, GPU-kernel generation, emulator construction, and autonomous nanoGPT speedruns. On Factorio, we find refinement allows for continuous technology progression and dedicated subagents enable parallelized work. Code is available at https://github.com/PrimeIntellect-ai/prime-agent.