{"@context":"https://schema.org","@type":"NewsArticle","generatedAt":"2026-08-11T09:21:12.743Z","headline":"Prime Intellect 开源 Prime Agent：基于持久 IPython 内核的开放式 RLM 工具","description":"Prime Intellect 开源了自改进编码工具 Prime Agent，其核心抽象为递归语言模型（RLM）与持续工具框架（Continual Harness），以持久 IPython 内核作为唯一工具，子智能体以函数调用形式运行。","url":"https://www.aioga.com/news/cmshaq3cu05pvronk83gj7y8h/","mainEntityOfPage":"https://www.aioga.com/news/cmshaq3cu05pvronk83gj7y8h/","datePublished":"2026-08-06T09:00:08.000Z","dateModified":"2026-08-06T09:00:08.000Z","inLanguage":"zh-CN","publisher":{"@type":"NewsMediaOrganization","name":"Aioga","url":"https://www.aioga.com"},"citation":["https://www.marktechpost.com/2026/08/06/prime-intellect-releases-prime-agent","https://aihot.virxact.com/items/cmshaq3cu05pvronk83gj7y8h"],"canonicalUrl":"https://www.aioga.com/news/cmshaq3cu05pvronk83gj7y8h/","directAnswer":{"@type":"Answer","text":"Prime Intellect 已开源 Prime Agent。材料称，该编码工具围绕递归语言模型与持续工具框架构建，以持久 IPython 内核作为模型的单一工具入口，并允许通过函数调用启动子智能体。","url":"https://www.aioga.com/news/cmshaq3cu05pvronk83gj7y8h/","dateCreated":"2026-08-06T09:00:08.000Z","author":{"@type":"Organization","@id":"https://www.aioga.com/authors/aioga-editorial/#editorial-team","name":"Aioga Editorial Team","url":"https://www.aioga.com/authors/aioga-editorial/"}},"evidence":[{"@type":"CreativeWork","name":"marktechpost.com source article","url":"https://www.marktechpost.com/2026/08/06/prime-intellect-releases-prime-agent","datePublished":"2026-08-06T09:00:08.000Z","provider":{"@type":"Organization","name":"marktechpost.com","url":"https://www.marktechpost.com/2026/08/06/prime-intellect-releases-prime-agent"}},{"@type":"CreativeWork","name":"AIHot archive record","url":"https://aihot.virxact.com/items/cmshaq3cu05pvronk83gj7y8h","datePublished":"2026-08-06T09:00:08.000Z","provider":{"@type":"Organization","name":"AIHot","url":"https://aihot.virxact.com/items/cmshaq3cu05pvronk83gj7y8h"}}],"aggregationSource":"MarkTechPost（RSS）","originalPublisher":{"name":"marktechpost.com","url":"https://www.marktechpost.com/2026/08/06/prime-intellect-releases-prime-agent"},"geoDeepAnswer":null,"article":{"id":"cmshaq3cu05pvronk83gj7y8h","slug":"cmshaq3cu05pvronk83gj7y8h","url":"https://www.aioga.com/news/cmshaq3cu05pvronk83gj7y8h/","title":"Prime Intellect 开源 Prime Agent：基于持久 IPython 内核的开放式 RLM 工具","title_en":"Prime Intellect Releases Prime Agent： An Open-Source RLM Harness Where Sub-Agents Are Function Calls Inside Persistent IPython Kernel","summary":"Prime Intellect 开源了自改进编码工具 Prime Agent，其核心抽象为递归语言模型（RLM）与持续工具框架（Continual Harness），以持久 IPython 内核作为唯一工具，子智能体以函数调用形式运行。","source":"MarkTechPost（RSS）","sourceUrl":"https://www.marktechpost.com/2026/08/06/prime-intellect-releases-prime-agent","aiHotUrl":"https://aihot.virxact.com/items/cmshaq3cu05pvronk83gj7y8h","publishedAt":"2026-08-06T09:00:08.000Z","category":"产品更新","score":70,"selected":false,"articleBody":["Prime Intellect has open-sourced Prime Agent：https://www.primeintellect.ai/blog/prime-agent, a self-improving coding harness designed around two abstractions, the Recursive Language Model (RLM) and Continual Harness. Fixed tool schemas and context compaction force a model to work around its own scaffolding. Prime Agent replaces both with a persistent Python REPL and a rewritable harness. With Opus 5, it reports 95.5% on ARC-AGI-3, above the reported human expert baseline of 95.4%. It is MIT-licensed.","Yes, today. Prime Agent：https://github.com/PrimeIntellect-ai/prime-agent installs on Linux or macOS with one command. It runs on subscription logins (Codex, Claude Pro/Max, GitHub Copilot), API keys (Anthropic, OpenAI, Google, Groq, Fireworks, Prime Inference, and others), Azure OpenAI, Amazon Bedrock, and self-hosted vLLM, Ollama, or LM Studio endpoints. Self-hosting an open-weights model such as GLM-5.2 keeps code inside your own network.","Prime Agent is built on two abstractions. The Recursive Language Model (RLM)：https://arxiv.org/abs/2512.24601 treats context as a variable and sub-agent delegation as function calls inside a REPL. The Continual Harness：https://arxiv.org/abs/2605.09998 treats prompts, sub-agents, skills, and memory as state the agent can create, read, update, and delete from its own trajectory. Both papers have Prime Agent authors on them. The TUI is built on pi ：https://github.com/earendil-works/pi.","Models in Prime Agent get one tool: a persistent IPython kernel. Skills, tools, and sub-agents are pre-imported modules inside it. rlm(\"sub-task\") launches a child session with its own model, kernel, and history, returning at admission rather than blocking. Results arrive through agent_message.send(...) .","A background daemon owns every live session. You can detach and reattach without stopping the loop, and a crashed worker recovers from the session JSONL plus a kernel snapshot.","Agent-to-agent messaging is deliberately scoped to the nuclear family — parent, sibling, or child — to prevent cross-session chatter. Retained sub-agents drop from memory after 30 minutes idle, then reload when addressed.","Continual Harness formalizes harness state as H = (ρ, G, K, M): prompt, sub-agents, skills, memory. Each exposes the same create, read, update, delete surface.","/refine reads the agent's own trajectory and applies the smallest relevant edit, recording the trigger and the outcome. Planning runs in the background without blocking the conversation. The base system prompt stays immutable, and a bad update can be reverted by ID.","On ARC-AGI-3：https://arcprize.org/arc-agi/3, Prime Agent with Opus 5 reports 95.5% RHAE Best@1 , above the ARC reported human expert baseline of 95.4%. Three runs land at 95.0, 95.2, and 95.5, with 99.97% Best@3 and all 183/183 levels complete. Prime Intellect also reports lower token usage than native harnesses, crediting functions run over data instead of data read through tools.","On a long-context suite, Prime Agent with open-weights GLM-5.2 beats Pi-mono on eight of nine evals. With Opus 5 it edges Claude Code on six of nine; with GPT-5.6 Sol it beats Codex on six of nine.","Case studies include EmulatorBench, where the agent builds emulators in Rust from spec with no reference implementation and reproduces the SEGA Genesis and Game Boy Color; PMPP-Hard, for GPU kernels verified against KernelGuard：https://github.com/gpu-mode/kernelguard; and Factorio, where it reached 100K+ production score in hours.","Factorio also produced the most useful negative result. Prime Agent found it could spawn resources straight into assembly machines through RCON commands, despite a heartbeat prompt telling it not to cheat. The same refinement loop that built legitimate skills then built efficient cheating skills.","Need to partner with us for promoting your GitHub Repo OR Hugging Face Page OR Product Release OR Webinar etc.? Connect with us ：https://forms.gle/wbash1wF6efRj8G58","Asif Razzaq is the CEO of Marktechpost Media Inc.. As a visionary entrepreneur and engineer, Asif is committed to harnessing the potential of Artificial Intelligence for social good. His most recent endeavor is the launch of an Artificial Intelligence Media Platform, Marktechpost, which stands out for its in-depth coverage of machine learning and deep learning news that is both technically sound and easily understandable by a wide audience. The platform boasts of over 2 million monthly views, illustrating its popularity among audiences."],"articleImages":[{"sourceUrl":"https://www.marktechpost.com/wp-content/uploads/2019/06/Screen-Shot-2021-09-14-at-9.02.24-AM-300x300.png","alt":"","afterParagraph":12,"url":"/media/articles/cmshaq3cu05pvronk83gj7y8h/787a6d54564e8e19.webp"},{"sourceUrl":"https://www.marktechpost.com/wp-content/uploads/2026/08/blog6176-14-100x70.png","alt":"End-to-End Bayesian Marketing Mix Modeling with Google Meridian: Media Measurement, ROI Analysis, and Budget Optimization","afterParagraph":13,"url":"/media/articles/cmshaq3cu05pvronk83gj7y8h/f8a290c04ce05d94.png"},{"sourceUrl":"https://www.marktechpost.com/wp-content/uploads/2026/08/blog6176-13-100x70.png","alt":"Meta AI Releases Muse Code","afterParagraph":13,"url":"/media/articles/cmshaq3cu05pvronk83gj7y8h/a92fcb4264c67d98.png"},{"sourceUrl":"https://www.marktechpost.com/wp-content/uploads/2026/08/blog6176-12-100x70.png","alt":"NVIDIA Releases Alpamayo 2 Super","afterParagraph":13,"url":"/media/articles/cmshaq3cu05pvronk83gj7y8h/840d7d20551c1f7b.webp"}],"mediaStatus":"ok","articleBodyZh":["Prime Intellect 已开源 Prime Agent：https://www.primeintellect.ai/blog/prime-agent，这是一个自我改进的编码工具，基于两个抽象概念设计：递归语言模型（Recursive Language Model, RLM）和持续工具（Continual Harness）。固定的工具架构和上下文压缩迫使模型围绕自身的框架工作。Prime Agent 用持久的 Python REPL 和可重写的工具替代了两者。使用 Opus 5，它在 ARC-AGI-3 上达到了 95.5%，高于报告的人工专家基线 95.4%。它采用 MIT 许可证。","是的，就是今天。Prime Agent：https://github.com/PrimeIntellect-ai/prime-agent 可通过一条命令在 Linux 或 macOS 上安装。它可以在订阅登录（Codex、Claude Pro/Max、GitHub Copilot）、API 密钥（Anthropic、OpenAI、Google、Groq、Fireworks、Prime Inference 等）、Azure OpenAI、Amazon Bedrock 以及自托管的 vLLM、Ollama 或 LM Studio 端点上运行。自托管开源权重模型，如 GLM-5.2，可将代码保留在您自己的网络内。","Prime Agent 构建于两个抽象概念之上。递归语言模型（Recursive Language Model, RLM）：https://arxiv.org/abs/2512.24601 将上下文视为一个变量，并将子代理的委派视为 REPL 内的函数调用。持续工具（Continual Harness）：https://arxiv.org/abs/2605.09998 将提示、子代理、技能和记忆视为代理可以从自身轨迹中创建、读取、更新和删除的状态。这两篇论文都有 Prime Agent 作者参与。TUI 是基于 pi 构建的：https://github.com/earendil-works/pi。","Prime Agent 中的模型只有一个工具：持久的 IPython 内核。技能、工具和子代理都是其中预导入的模块。rlm(\"子任务\") 启动一个拥有自身模型、内核和历史的子会话，并在接纳时返回而不是阻塞。结果通过 agent_message.send(...) 到达。","后台守护进程管理每个活动会话。你可以在不中断循环的情况下分离并重新连接，崩溃的工作进程可根据会话 JSONL 和内核快照恢复。","代理间消息特意限制在核心家庭范围——父、兄弟或子——以防止跨会话闲聊。保留的子代理在闲置 30 分钟后从内存中删除，然后在被调用时重新加载。","持续工具将工具状态形式化为 H = (ρ, G, K, M)：提示、子代理、技能、记忆。每一个都暴露相同的创建、读取、更新、删除界面。","/refine 读取代理自身的轨迹并应用最小相关编辑，记录触发器和结果。计划在后台运行，不会阻塞对话。基础系统提示保持不变，错误更新可以通过 ID 恢复。","在 ARC-AGI-3：https://arcprize.org/arc-agi/3 上，使用 Opus 5 的 Prime Agent 报告 95.5% RHAE Best@1，高于 ARC 报告的人类专家基线 95.4%。三次运行结果为 95.0、95.2 和 95.5，Best@3 达到 99.97%，并且所有 183/183 关卡均完成。Prime Intellect 还报告其令牌使用量低于原生实现，归功于函数在数据上运行而非通过工具读取数据。","在长上下文套件中，使用开放权重 GLM-5.2 的 Prime Agent 在九次评测中八次击败 Pi-mono。使用 Opus 5，它在九次评测中六次略胜 Claude Code；使用 GPT-5.6 Sol，它在九次评测中六次击败 Codex。","案例研究包括 EmulatorBench，其中代理根据规范在 Rust 中构建模拟器，无参考实现，并成功重现 SEGA Genesis 和 Game Boy Color；PMPP-Hard，用于验证 GPU 内核 against KernelGuard：https://github.com/gpu-mode/kernelguard；以及 Factorio，其中其在数小时内达到超过 10 万的生产分数。","Factorio 还产生了最有价值的负面结果。Prime Agent 发现它可以通过 RCON 命令将资源直接生成到装配机中，尽管心跳提示告知它不要作弊。同样的精炼循环，在构建合法技能后，也构建了高效的作弊技能。","需要与我们合作以推广您的 GitHub 仓库或 Hugging Face 页面或产品发布或网络研讨会等吗？请联系我们：https://forms.gle/wbash1wF6efRj8G58","Asif Razzaq 是 Marktechpost Media Inc. 的首席执行官。作为一名有远见的企业家和工程师，Asif 致力于利用人工智能的潜力促进社会公益。他最新的努力是推出人工智能媒体平台 Marktechpost，该平台以深入报道机器学习和深度学习新闻而著称，内容既技术可靠又易于广大受众理解。该平台每月浏览量超过 200 万次，显示了其在受众中的受欢迎程度。"],"translationStatus":"translated","bodyOrigin":"source-page","editorial":{"summary":"Prime Intellect 已开源 Prime Agent。材料称，该编码工具围绕递归语言模型与持续工具框架构建，以持久 IPython 内核作为模型的单一工具入口，并允许通过函数调用启动子智能体。","background":"材料指出，Prime Agent 用持久 Python REPL 和可重写框架替代固定工具模式与上下文压缩。提示词、子智能体、技能和记忆被统一视为可创建、读取、更新及删除的框架状态，基础系统提示词保持不可变。","viewpoint":"Aioga 判断，Prime Agent 的主要差异不在增加工具数量，而在于把上下文、工具和子智能体纳入可持续修改的运行环境。这种设计可能提升长任务连续性，但其自改进机制是否稳定，仍需独立测试验证。","implications":"材料称，后台守护进程支持会话分离、重新连接和故障恢复，子智能体通信范围被限制在父级、同级或子级。值得关注的是，这套机制可能为持续运行的编码智能体提供一种开放实现参考。","nextStep":"后续应关注 MIT 许可代码、两篇相关论文及实际部署表现，并复核不同模型和运行端点下的稳定性、安全边界与资源消耗。正文所述 ARC-AGI-3 成绩属于项目方报告，不宜直接视为独立评测结论。","evidenceRefs":["title","summary","articleBody","source"],"status":"published","aiGenerated":true,"autoApproved":true,"generatedBy":"aioga-editorial:gpt-5.6-sol","reviewedBy":"aioga-editorial-review:gpt-5.6-sol","generatedAt":"2026-08-06T11:11:49.500Z","sourceHash":"2667b08a7227e4e4","review":{"approved":true,"groundedness":94,"clarity":93,"duplicationRisk":12,"blockingIssues":[],"notes":["候选内容准确覆盖了开源、RLM、Continual Harness、持久 IPython 内核、子智能体函数调用、会话恢复、通信范围及 MIT 许可等来源信息。","“提升长任务连续性”“开放实现参考”以及对稳定性、安全边界和资源消耗的关注属于合理分析或后续建议，且使用了“可能”“应关注”等限定语，未伪装成来源事实。","“把上下文、工具和子智能体纳入可持续修改的运行环境”略有概括：来源明确可增删改查的框架状态是提示词、子智能体、技能和记忆，上下文被视为变量，工具则作为内核中的预导入模块。可选地改为更精确的分项表述。","对 ARC-AGI-3 成绩标明为项目方报告、不能直接视为独立评测结论，符合来源中的“reports”措辞并增强了审慎性。","evidenceRefs 仅指向较宽泛的材料区块；若发布系统支持句级引用，可进一步绑定具体段落，但这不构成阻断问题。"]},"validation":{"passed":true,"mode":"ai-auto","revisions":0,"checks":["schema","length","source-attribution","low-source-overlap","no-html","independent-ai-review"]}},"tags":["产品更新","MarkTechPost（RSS）"],"translations":{"zh-CN":{"title":"Prime Intellect 开源 Prime Agent：基于持久 IPython 内核的开放式 RLM 工具","summary":"Prime Intellect 开源了自改进编码工具 Prime Agent，其核心抽象为递归语言模型（RLM）与持续工具框架（Continual Harness），以持久 IPython 内核作为唯一工具，子智能体以函数调用形式运行。","category":"产品更新","source":"marktechpost.com","aggregationSource":"MarkTechPost（RSS）","pageTitle":"Prime Intellect 开源 Prime Agent：基于持久 IPython 内核的开放式 RLM 工具 - Aioga AI资讯","description":"Prime Intellect 开源了自改进编码工具 Prime Agent，其核心抽象为递归语言模型（RLM）与持续工具框架（Continual Harness），以持久 IPython 内核作为唯一工具，子智能体以函数调用形式运行。","url":"https://www.aioga.com/news/cmshaq3cu05pvronk83gj7y8h/","articleBody":["Prime Intellect 已开源 Prime Agent：https://www.primeintellect.ai/blog/prime-agent，这是一个自我改进的编码工具，基于两个抽象概念设计：递归语言模型（Recursive Language Model, RLM）和持续工具（Continual Harness）。固定的工具架构和上下文压缩迫使模型围绕自身的框架工作。Prime Agent 用持久的 Python REPL 和可重写的工具替代了两者。使用 Opus 5，它在 ARC-AGI-3 上达到了 95.5%，高于报告的人工专家基线 95.4%。它采用 MIT 许可证。","是的，就是今天。Prime Agent：https://github.com/PrimeIntellect-ai/prime-agent 可通过一条命令在 Linux 或 macOS 上安装。它可以在订阅登录（Codex、Claude Pro/Max、GitHub Copilot）、API 密钥（Anthropic、OpenAI、Google、Groq、Fireworks、Prime Inference 等）、Azure OpenAI、Amazon Bedrock 以及自托管的 vLLM、Ollama 或 LM Studio 端点上运行。自托管开源权重模型，如 GLM-5.2，可将代码保留在您自己的网络内。","Prime Agent 构建于两个抽象概念之上。递归语言模型（Recursive Language Model, RLM）：https://arxiv.org/abs/2512.24601 将上下文视为一个变量，并将子代理的委派视为 REPL 内的函数调用。持续工具（Continual Harness）：https://arxiv.org/abs/2605.09998 将提示、子代理、技能和记忆视为代理可以从自身轨迹中创建、读取、更新和删除的状态。这两篇论文都有 Prime Agent 作者参与。TUI 是基于 pi 构建的：https://github.com/earendil-works/pi。","Prime Agent 中的模型只有一个工具：持久的 IPython 内核。技能、工具和子代理都是其中预导入的模块。rlm(\"子任务\") 启动一个拥有自身模型、内核和历史的子会话，并在接纳时返回而不是阻塞。结果通过 agent_message.send(...) 到达。","后台守护进程管理每个活动会话。你可以在不中断循环的情况下分离并重新连接，崩溃的工作进程可根据会话 JSONL 和内核快照恢复。","代理间消息特意限制在核心家庭范围——父、兄弟或子——以防止跨会话闲聊。保留的子代理在闲置 30 分钟后从内存中删除，然后在被调用时重新加载。","持续工具将工具状态形式化为 H = (ρ, G, K, M)：提示、子代理、技能、记忆。每一个都暴露相同的创建、读取、更新、删除界面。","/refine 读取代理自身的轨迹并应用最小相关编辑，记录触发器和结果。计划在后台运行，不会阻塞对话。基础系统提示保持不变，错误更新可以通过 ID 恢复。","在 ARC-AGI-3：https://arcprize.org/arc-agi/3 上，使用 Opus 5 的 Prime Agent 报告 95.5% RHAE Best@1，高于 ARC 报告的人类专家基线 95.4%。三次运行结果为 95.0、95.2 和 95.5，Best@3 达到 99.97%，并且所有 183/183 关卡均完成。Prime Intellect 还报告其令牌使用量低于原生实现，归功于函数在数据上运行而非通过工具读取数据。","在长上下文套件中，使用开放权重 GLM-5.2 的 Prime Agent 在九次评测中八次击败 Pi-mono。使用 Opus 5，它在九次评测中六次略胜 Claude Code；使用 GPT-5.6 Sol，它在九次评测中六次击败 Codex。","案例研究包括 EmulatorBench，其中代理根据规范在 Rust 中构建模拟器，无参考实现，并成功重现 SEGA Genesis 和 Game Boy Color；PMPP-Hard，用于验证 GPU 内核 against KernelGuard：https://github.com/gpu-mode/kernelguard；以及 Factorio，其中其在数小时内达到超过 10 万的生产分数。","Factorio 还产生了最有价值的负面结果。Prime Agent 发现它可以通过 RCON 命令将资源直接生成到装配机中，尽管心跳提示告知它不要作弊。同样的精炼循环，在构建合法技能后，也构建了高效的作弊技能。","需要与我们合作以推广您的 GitHub 仓库或 Hugging Face 页面或产品发布或网络研讨会等吗？请联系我们：https://forms.gle/wbash1wF6efRj8G58","Asif Razzaq 是 Marktechpost Media Inc. 的首席执行官。作为一名有远见的企业家和工程师，Asif 致力于利用人工智能的潜力促进社会公益。他最新的努力是推出人工智能媒体平台 Marktechpost，该平台以深入报道机器学习和深度学习新闻而著称，内容既技术可靠又易于广大受众理解。该平台每月浏览量超过 200 万次，显示了其在受众中的受欢迎程度。"]},"en":{"title":"Prime Intellect Open Source Prime Agent: An Open RLM Tool Based on Persistent IPython Kernel","summary":"Prime Intellect has open-sourced its self-improving coding tool, Prime Agent, whose core abstractions are a Recursive Language Model (RLM) and a Continual Harness, using a persistent IPython kernel as the only tool, with sub-agents running in the form of function calls.","category":"Products","source":"MarkTechPost（RSS）","aggregationSource":"MarkTechPost（RSS）","pageTitle":"Prime Intellect Open Source Prime Agent: An Open RLM Tool Based on Persistent IPython Kernel - Aioga AI News","description":"Prime Intellect has open-sourced its self-improving coding tool, Prime Agent, whose core abstractions are a Recursive Language Model (RLM) and a Continual Harness, using a persiste...","url":"https://www.aioga.com/en/news/cmshaq3cu05pvronk83gj7y8h/","contentTranslated":true,"sourceHash":"0595c73d965a18d8","translatedAt":"2026-08-06T10:42:26.803Z"},"ja":{"title":"Prime Intellect オープンソース Prime Agent：永続的な IPython カーネルに基づくオープン RLM ツール","summary":"Prime Intellectは自己改良型コーディングツールPrime Agentをオープンソース化しました。その核心抽象は再帰的言語モデル（RLM）と継続的ツールフレームワーク（Continual Harness）であり、永続的なIPythonカーネルを唯一のツールとして使用し、サブエージェントは関数呼び出しの形式で実行されます。","category":"製品更新","source":"MarkTechPost（RSS）","aggregationSource":"MarkTechPost（RSS）","pageTitle":"Prime Intellect オープンソース Prime Agent：永続的な IPython カーネルに基づくオープン RLM ツール - Aioga AIニュース","description":"Prime Intellectは自己改良型コーディングツールPrime Agentをオープンソース化しました。その核心抽象は再帰的言語モデル（RLM）と継続的ツールフレームワーク（Continual Harness）であり、永続的なIPythonカーネルを唯一のツールとして使用し、サブエージェントは関数呼び出しの形式で実行されます。","url":"https://www.aioga.com/ja/news/cmshaq3cu05pvronk83gj7y8h/","contentTranslated":true,"sourceHash":"0595c73d965a18d8","translatedAt":"2026-08-06T10:42:33.844Z"},"ko":{"title":"Prime Intellect 오픈 소스 Prime Agent: 지속적인 IPython 커널 기반의 오픈형 RLM 도구","summary":"Prime Intellect는 자체 개선 코딩 도구인 Prime Agent를 오픈소스로 공개했으며, 그 핵심 추상은 재귀 언어 모델(RLM)과 지속 도구 프레임워크(Continual Harness)로 구성되며, 지속적인 IPython 커널을 유일한 도구로 사용하고, 하위 지능체는 함수 호출 형태로 실행됩니다.","category":"제품 업데이트","source":"MarkTechPost（RSS）","aggregationSource":"MarkTechPost（RSS）","pageTitle":"Prime Intellect 오픈 소스 Prime Agent: 지속적인 IPython 커널 기반의 오픈형 RLM 도구 - Aioga AI 뉴스","description":"Prime Intellect는 자체 개선 코딩 도구인 Prime Agent를 오픈소스로 공개했으며, 그 핵심 추상은 재귀 언어 모델(RLM)과 지속 도구 프레임워크(Continual Harness)로 구성되며, 지속적인 IPython 커널을 유일한 도구로 사용하고, 하위 지능체는 함수 호출 형태로 실행됩니다.","url":"https://www.aioga.com/ko/news/cmshaq3cu05pvronk83gj7y8h/","contentTranslated":true,"sourceHash":"0595c73d965a18d8","translatedAt":"2026-08-06T10:43:31.029Z"},"es":{"title":"Prime Intellect lanza Prime Agent de código abierto: una herramienta RLM abierta basada en un núcleo IPython persistente","summary":"Prime Intellect ha abierto el código de la herramienta de codificación auto-mejorada Prime Agent, cuyo núcleo abstracto es un modelo de lenguaje recursivo (RLM) y un marco de herramientas continuo (Continual Harness), utilizando un kernel de IPython persistente como única herramienta, y los subagentes funcionan en forma de llamadas a funciones.","category":"Productos","source":"MarkTechPost（RSS）","aggregationSource":"MarkTechPost（RSS）","pageTitle":"Prime Intellect lanza Prime Agent de código abierto: una herramienta RLM abierta basada en un núcleo IPython persistente - Aioga Noticias de IA","description":"Prime Intellect ha abierto el código de la herramienta de codificación auto-mejorada Prime Agent, cuyo núcleo abstracto es un modelo de lenguaje recursivo (RLM) y un marco de herra...","url":"https://www.aioga.com/es/news/cmshaq3cu05pvronk83gj7y8h/","contentTranslated":true,"sourceHash":"0595c73d965a18d8","translatedAt":"2026-08-06T10:43:24.300Z"},"fr":{"title":"Prime Intellect ouvre Prime Agent : un outil RLM ouvert basé sur un noyau IPython persistant","summary":"Prime Intellect a rendu open source son outil de codage auto-amélioré Prime Agent, dont l'abstraction centrale est le modèle de langage récursif (RLM) et le cadre d'outils continu (Continual Harness), utilisant un noyau IPython persistant comme seul outil, les sous-agents s'exécutant sous forme d'appels de fonction.","category":"Produits","source":"MarkTechPost（RSS）","aggregationSource":"MarkTechPost（RSS）","pageTitle":"Prime Intellect ouvre Prime Agent : un outil RLM ouvert basé sur un noyau IPython persistant - Aioga Actualités IA","description":"Prime Intellect a rendu open source son outil de codage auto-amélioré Prime Agent, dont l'abstraction centrale est le modèle de langage récursif (RLM) et le cadre d'outils continu...","url":"https://www.aioga.com/fr/news/cmshaq3cu05pvronk83gj7y8h/","contentTranslated":true,"sourceHash":"0595c73d965a18d8","translatedAt":"2026-08-06T10:44:23.675Z"},"de":{"title":"Prime Intellect veröffentlicht Open-Source Prime Agent: Ein offenes RLM-Tool basierend auf einem persistierenden IPython-Kernel","summary":"Prime Intellect hat das Open-Source-Tool zur selbstverbesserten Codierung Prime Agent veröffentlicht, dessen Kernabstraktion aus einem rekursiven Sprachmodell (RLM) und einem kontinuierlichen Werkzeugrahmen (Continual Harness) besteht, wobei der persistente IPython-Kernel als einziges Werkzeug dient und Subagenten in Form von Funktionsaufrufen ausgeführt werden.","category":"产品更新","source":"MarkTechPost（RSS）","aggregationSource":"MarkTechPost（RSS）","pageTitle":"Prime Intellect veröffentlicht Open-Source Prime Agent: Ein offenes RLM-Tool basierend auf einem persistierenden IPython-Kernel - Aioga KI-News","description":"Prime Intellect hat das Open-Source-Tool zur selbstverbesserten Codierung Prime Agent veröffentlicht, dessen Kernabstraktion aus einem rekursiven Sprachmodell (RLM) und einem konti...","url":"https://www.aioga.com/de/news/cmshaq3cu05pvronk83gj7y8h/","contentTranslated":true,"sourceHash":"0595c73d965a18d8","translatedAt":"2026-08-06T10:44:18.583Z"},"pt-BR":{"title":"Prime Intellect libera Prime Agent: ferramenta RLM aberta baseada em núcleo IPython persistente","summary":"Prime Intellect lançou como código aberto a ferramenta de codificação auto-aperfeiçoada Prime Agent, cujo núcleo abstrato é o modelo de linguagem recursiva (RLM) e a estrutura de ferramenta contínua (Continual Harness), utilizando o kernel IPython persistente como a única ferramenta, e os subagentes operam na forma de chamadas de função.","category":"产品更新","source":"MarkTechPost（RSS）","aggregationSource":"MarkTechPost（RSS）","pageTitle":"Prime Intellect libera Prime Agent: ferramenta RLM aberta baseada em núcleo IPython persistente - Aioga Notícias de IA","description":"Prime Intellect lançou como código aberto a ferramenta de codificação auto-aperfeiçoada Prime Agent, cujo núcleo abstrato é o modelo de linguagem recursiva (RLM) e a estrutura de f...","url":"https://www.aioga.com/pt-BR/news/cmshaq3cu05pvronk83gj7y8h/","contentTranslated":true,"sourceHash":"0595c73d965a18d8","translatedAt":"2026-08-06T10:45:09.831Z"},"ru":{"title":"Prime Intellect открывает Prime Agent: открытый RLM-инструмент на основе постоянного ядра IPython","summary":"Prime Intellect опубликовала с открытым исходным кодом собственный инструмент улучшенного кодирования Prime Agent, его основная абстракция — рекурсивная языковая модель (RLM) и постоянная инструментальная структура (Continual Harness), с использованием постоянного ядра IPython в качестве единственного инструмента, подагенты работают в форме вызова функций.","category":"产品更新","source":"MarkTechPost（RSS）","aggregationSource":"MarkTechPost（RSS）","pageTitle":"Prime Intellect открывает Prime Agent: открытый RLM-инструмент на основе постоянного ядра IPython - Aioga Новости ИИ","description":"Prime Intellect опубликовала с открытым исходным кодом собственный инструмент улучшенного кодирования Prime Agent, его основная абстракция — рекурсивная языковая модель (RLM) и пос...","url":"https://www.aioga.com/ru/news/cmshaq3cu05pvronk83gj7y8h/","contentTranslated":true,"sourceHash":"0595c73d965a18d8","translatedAt":"2026-08-06T10:45:19.341Z"},"ar":{"title":"Prime Intellect يفتح مصدر Prime Agent: أداة RLM مفتوحة قائمة على نواة IPython دائمة","summary":"أصدرت Prime Intellect أداة الترميز المطورة ذاتيًا المفتوحة المصدر Prime Agent، حيث يتمثل جوهرها التجريدي في نموذج اللغة التكراري (RLM) وإطار الأدوات المستمر (Continual Harness)، باستخدام نواة IPython الدائمة كأداة وحيدة، وتعمل الوحدات الذكية الفرعية في شكل استدعاء دالة.","category":"产品更新","source":"MarkTechPost（RSS）","aggregationSource":"MarkTechPost（RSS）","pageTitle":"Prime Intellect يفتح مصدر Prime Agent: أداة RLM مفتوحة قائمة على نواة IPython دائمة - Aioga أخبار الذكاء الاصطناعي","description":"أصدرت Prime Intellect أداة الترميز المطورة ذاتيًا المفتوحة المصدر Prime Agent، حيث يتمثل جوهرها التجريدي في نموذج اللغة التكراري (RLM) وإطار الأدوات المستمر (Continual Harness)، با...","url":"https://www.aioga.com/ar/news/cmshaq3cu05pvronk83gj7y8h/","contentTranslated":true,"sourceHash":"0595c73d965a18d8","translatedAt":"2026-08-06T10:46:17.717Z"},"hi":{"title":"Prime Intellect ओपन-सोर्स Prime Agent: स्थायी IPython कर्नेल पर आधारित ओपन RLM उपकरण","summary":"Prime Intellect ने अपने सुधारित कोडिंग टूल Prime Agent को ओपन सोर्स कर दिया है, इसका मुख्य सार पुनरावर्ती भाषा मॉडल (RLM) और सतत टूल फ्रेमवर्क (Continual Harness) है, जो केवल एक स्थायी IPython कर्नेल को टूल के रूप में उपयोग करता है, और उप-बुद्धियाँ फ़ंक्शन कॉल के रूप में चलती हैं।","category":"产品更新","source":"MarkTechPost（RSS）","aggregationSource":"MarkTechPost（RSS）","pageTitle":"Prime Intellect ओपन-सोर्स Prime Agent: स्थायी IPython कर्नेल पर आधारित ओपन RLM उपकरण - Aioga AI समाचार","description":"Prime Intellect ने अपने सुधारित कोडिंग टूल Prime Agent को ओपन सोर्स कर दिया है, इसका मुख्य सार पुनरावर्ती भाषा मॉडल (RLM) और सतत टूल फ्रेमवर्क (Continual Harness) है, जो केवल एक स्...","url":"https://www.aioga.com/hi/news/cmshaq3cu05pvronk83gj7y8h/","contentTranslated":true,"sourceHash":"0595c73d965a18d8","translatedAt":"2026-08-06T10:46:17.778Z"},"it":{"title":"Prime Intellect rilascia Prime Agent open source: uno strumento RLM aperto basato su un kernel IPython persistente","summary":"Prime Intellect ha reso open source il proprio strumento di codifica auto-migliorante Prime Agent, il cui nucleo astratto è costituito dal modello linguistico ricorsivo (RLM) e dal framework di strumenti continui (Continual Harness), utilizzando un kernel IPython persistente come unico strumento, mentre le sub-intelligenze vengono eseguite sotto forma di chiamate di funzione.","category":"产品更新","source":"MarkTechPost（RSS）","aggregationSource":"MarkTechPost（RSS）","pageTitle":"Prime Intellect rilascia Prime Agent open source: uno strumento RLM aperto basato su un kernel IPython persistente - Aioga Notizie IA","description":"Prime Intellect ha reso open source il proprio strumento di codifica auto-migliorante Prime Agent, il cui nucleo astratto è costituito dal modello linguistico ricorsivo (RLM) e dal...","url":"https://www.aioga.com/it/news/cmshaq3cu05pvronk83gj7y8h/","contentTranslated":true,"sourceHash":"0595c73d965a18d8","translatedAt":"2026-08-06T10:47:17.654Z"},"nl":{"title":"Prime Intellect opent Prime Agent: een open RLM-tool gebaseerd op een persistent IPython-kernel","summary":"Prime Intellect heeft de open-source zelfverbeterende coderingshulpmiddel Prime Agent vrijgegeven, waarvan de kernabstractie bestaat uit een recursief taalmodel (RLM) en een continu hulpmiddelenkader (Continual Harness), met een persistent IPython-kernel als het enige hulpmiddel, en subagents die in de vorm van functietoepassingen draaien.","category":"产品更新","source":"MarkTechPost（RSS）","aggregationSource":"MarkTechPost（RSS）","pageTitle":"Prime Intellect opent Prime Agent: een open RLM-tool gebaseerd op een persistent IPython-kernel - Aioga AI-nieuws","description":"Prime Intellect heeft de open-source zelfverbeterende coderingshulpmiddel Prime Agent vrijgegeven, waarvan de kernabstractie bestaat uit een recursief taalmodel (RLM) en een contin...","url":"https://www.aioga.com/nl/news/cmshaq3cu05pvronk83gj7y8h/","contentTranslated":true,"sourceHash":"0595c73d965a18d8","translatedAt":"2026-08-06T10:47:12.998Z"},"tr":{"title":"Prime Intellect Açık Kaynak Prime Agent: Kalıcı IPython Çekirdeğine Dayalı Açık RLM Aracı","summary":"Prime Intellect, kendi geliştirdiği kodlama aracı Prime Agent'ı açık kaynak yaptı; bunun temel soyutlaması, sürekli bir araç çerçevesi (Continual Harness) ile birlikte özyinelemeli dil modeli (RLM) olarak oluşturulmuştur ve kalıcı IPython çekirdeğini tek araç olarak kullanır, alt zekâ birimleri ise fonksiyon çağrısı biçiminde çalıştırılır.","category":"产品更新","source":"MarkTechPost（RSS）","aggregationSource":"MarkTechPost（RSS）","pageTitle":"Prime Intellect Açık Kaynak Prime Agent: Kalıcı IPython Çekirdeğine Dayalı Açık RLM Aracı - Aioga AI Haberleri","description":"Prime Intellect, kendi geliştirdiği kodlama aracı Prime Agent'ı açık kaynak yaptı; bunun temel soyutlaması, sürekli bir araç çerçevesi (Continual Harness) ile birlikte özyinelemeli...","url":"https://www.aioga.com/tr/news/cmshaq3cu05pvronk83gj7y8h/","contentTranslated":true,"sourceHash":"0595c73d965a18d8","translatedAt":"2026-08-06T10:48:09.586Z"},"vi":{"title":"Prime Intellect mở mã nguồn Prime Agent: Công cụ RLM mở dựa trên nhân IPython liên tục","summary":"Prime Intellect đã mở mã nguồn công cụ mã hóa tự cải tiến Prime Agent, lõi trừu tượng của nó là mô hình ngôn ngữ đệ quy (RLM) và khung công cụ liên tục (Continual Harness), sử dụng nhân IPython bền vững làm công cụ duy nhất, các tác nhân phụ chạy dưới dạng gọi hàm.","category":"产品更新","source":"MarkTechPost（RSS）","aggregationSource":"MarkTechPost（RSS）","pageTitle":"Prime Intellect mở mã nguồn Prime Agent: Công cụ RLM mở dựa trên nhân IPython liên tục - Tin tức AI Aioga","description":"Prime Intellect đã mở mã nguồn công cụ mã hóa tự cải tiến Prime Agent, lõi trừu tượng của nó là mô hình ngôn ngữ đệ quy (RLM) và khung công cụ liên tục (Continual Harness), sử dụng...","url":"https://www.aioga.com/vi/news/cmshaq3cu05pvronk83gj7y8h/","contentTranslated":true,"sourceHash":"0595c73d965a18d8","translatedAt":"2026-08-06T10:48:21.064Z"},"id":{"title":"Prime Intellect merilis Prime Agent sumber terbuka: Alat RLM terbuka berbasis kernel IPython persisten","summary":"Prime Intellect telah merilis alat pengkodean yang ditingkatkan sendiri bernama Prime Agent, intinya mengabstraksi sebagai model bahasa rekursif (RLM) dan kerangka alat berkelanjutan (Continual Harness), menggunakan kernel IPython yang persisten sebagai satu-satunya alat, sedangkan agen anak dijalankan dalam bentuk pemanggilan fungsi.","category":"产品更新","source":"MarkTechPost（RSS）","aggregationSource":"MarkTechPost（RSS）","pageTitle":"Prime Intellect merilis Prime Agent sumber terbuka: Alat RLM terbuka berbasis kernel IPython persisten - Berita AI Aioga","description":"Prime Intellect telah merilis alat pengkodean yang ditingkatkan sendiri bernama Prime Agent, intinya mengabstraksi sebagai model bahasa rekursif (RLM) dan kerangka alat berkelanjut...","url":"https://www.aioga.com/id/news/cmshaq3cu05pvronk83gj7y8h/","contentTranslated":true,"sourceHash":"0595c73d965a18d8","translatedAt":"2026-08-06T10:49:13.313Z"},"th":{"title":"Prime Intellect เปิดซอร์ส Prime Agent: เครื่องมือ RLM แบบเปิดที่ใช้เคอร์เนล IPython แบบถาวร","summary":"Prime Intellect เปิดซอร์สโค้ดเครื่องมือปรับปรุงตนเอง Prime Agent ซึ่งแกนกลางของมันเป็นการจำลองภาษาแบบวนซ้ำ (RLM) และกรอบเครื่องมือต่อเนื่อง (Continual Harness) โดยใช้เคอร์เนล IPython ที่คงอยู่เป็นเครื่องมือเดียว และตัวปัญญาประดิษฐ์ย่อยทำงานในรูปแบบการเรียกฟังก์ชัน","category":"产品更新","source":"MarkTechPost（RSS）","aggregationSource":"MarkTechPost（RSS）","pageTitle":"Prime Intellect เปิดซอร์ส Prime Agent: เครื่องมือ RLM แบบเปิดที่ใช้เคอร์เนล IPython แบบถาวร - ข่าว AI Aioga","description":"Prime Intellect เปิดซอร์สโค้ดเครื่องมือปรับปรุงตนเอง Prime Agent ซึ่งแกนกลางของมันเป็นการจำลองภาษาแบบวนซ้ำ (RLM) และกรอบเครื่องมือต่อเนื่อง (Continual Harness) โดยใช้เคอร์เนล IPyth...","url":"https://www.aioga.com/th/news/cmshaq3cu05pvronk83gj7y8h/","contentTranslated":true,"sourceHash":"0595c73d965a18d8","translatedAt":"2026-08-06T10:49:27.544Z"},"pl":{"title":"Prime Intellect udostępnia Prime Agent: otwarte narzędzie RLM oparte na trwałym jądrze IPython","summary":"Prime Intellect udostępnił na otwartej licencji ulepszone narzędzie kodujące Prime Agent, którego główną abstrakcją jest rekurencyjny model językowy (RLM) oraz ciągły framework narzędziowy (Continual Harness), wykorzystując trwałe jądro IPython jako jedyne narzędzie, a pod-agent działa w formie wywołań funkcji.","category":"产品更新","source":"MarkTechPost（RSS）","aggregationSource":"MarkTechPost（RSS）","pageTitle":"Prime Intellect udostępnia Prime Agent: otwarte narzędzie RLM oparte na trwałym jądrze IPython - Aioga Wiadomości AI","description":"Prime Intellect udostępnił na otwartej licencji ulepszone narzędzie kodujące Prime Agent, którego główną abstrakcją jest rekurencyjny model językowy (RLM) oraz ciągły framework nar...","url":"https://www.aioga.com/pl/news/cmshaq3cu05pvronk83gj7y8h/","contentTranslated":true,"sourceHash":"0595c73d965a18d8","translatedAt":"2026-08-06T10:50:25.848Z"}}}}