Prime Intellect 开源 Prime Agent 编码框架
Prime Intellect Releases Prime Agent: An Open-Source RLM Harness Where Sub-Agents Are Function Calls Inside Persistent IPython Kernel
做 Agent 或编码工具的同学必看，Prime Agent 用持久内核和可重写框架替代固定工具，ARC-AGI-3 超人类基线，赶紧去 GitHub 看实现细节。
Prime Intellect has open-sourced 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.
Prime Intellect 已开源 Prime Agent，这是一个自我改进的编码工具，围绕两个抽象设计：递归语言模型（RLM）和持续工具（Continual Harness）。固定的工具模式和上下文压缩迫使模型在其自身脚手架的限制下工作。Prime Agent 用持久的 Python REPL 和可重写的工具替代了这两者。使用 Opus 5，它在 ARC-AGI-3 上报告了 95.5% 的准确率，高于报告的人类专家基线 95.4%。它采用 MIT 许可证。
Is it deployable
它是否可部署？
Yes, today. 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.
是的，今天就可以。Prime Agent 通过一条命令安装在 Linux 或 macOS 上。它支持订阅登录（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。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。
- Company level: Best fit is mid-size to large engineering orgs and AI labs that already run isolated CI containers. Prime Intellect states plainly that worker and kernel processes are not a security sandbox. Deployment therefore needs disposable clones or restricted environments. Solo developers can install it, but the payoff appears on multi-hour tasks.
- Industries: Developer tooling, semiconductor and HPC teams writing GPU kernels, simulation and gaming, quantitative research, and AI research labs.
- Applications: Overnight refactors behind a test gate, spec-driven builds from scratch, kernel optimization, long-horizon agent evaluation, and autoresearch.
- 公司层面：最适合中型到大型工程组织和已经运行隔离 CI 容器的 AI 实验室。Prime Intellect 明确表示，worker 和内核进程不是安全沙箱。因此，部署需要一次性克隆或受限环境。独立开发者可以安装它，但收益体现在多小时的长时间任务上。
- 行业：开发者工具、编写 GPU 内核的半导体和高性能计算团队、模拟和游戏、量化研究以及 AI 研究实验室。
- 应用：测试门禁后的隔夜重构、从零开始的规范驱动构建、内核优化、长周期智能体评估和自动研究。
What Prime Intellect shipped
Prime Intellect 发布了什么
Prime Agent is built on two abstractions. The Recursive Language Model (RLM) treats context as a variable and sub-agent delegation as function calls inside a REPL. The Continual Harness 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.
Prime Agent 基于两个抽象构建。递归语言模型（RLM）将上下文视为变量，将子智能体委派视为 REPL 中的函数调用。持续工具（Continual Harness）将提示、子智能体、技能和记忆视为智能体可以从其自身轨迹中创建、读取、更新和删除的状态。两篇论文都有 Prime Agent 的作者参与。TUI 基于 pi 构建。
Programmatic tool calling
程序化工具调用
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(...).
Prime Agent 中的模型获得一个工具：一个持久的 IPython 内核。技能、工具和子智能体是其中预导入的模块。rlm("子任务") 启动一个子会话，该会话有自己的模型、内核和历史记录，在准入时返回而不是阻塞。结果通过 agent_message.send(...) 到达。
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.
一个后台守护进程拥有每个活动会话。您可以在不停止循环的情况下分离并重新附加，崩溃的 worker 可以从会话 JSONL 和内核快照中恢复。
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.
代理间消息传递被刻意限定在核心家庭范围内——父母、兄弟姐妹或子女——以防止跨会话闲聊。保留的子代理在空闲30分钟后从内存中消失，在被调用时重新加载。
Self-improvement through /refine
通过 /refine 实现自我改进
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.
Continual Harness 将 harness 状态形式化为 H = (ρ, G, K, M)：提示、子代理、技能、记忆。每个都暴露相同的创建、读取、更新、删除接口。
/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.
/refine 读取代理自身的轨迹并应用最小的相关编辑，记录触发器和结果。规划在后台运行，不阻塞对话。基础系统提示保持不可变，错误的更新可以通过 ID 回滚。
Benchmarks
基准测试
On 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.
在 ARC-AGI-3 上，Prime Agent 搭配 Opus 5 报告了 95.5% 的 RHAE Best@1，高于 ARC 报告的人类专家基线 95.4%。三次运行分别达到 95.0、95.2 和 95.5，Best@3 为 99.97%，并完成了全部 183/183 个关卡。Prime Intellect 还报告了比原生 harness 更低的 token 使用量，这归功于在数据上运行函数而不是通过工具读取数据。
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.
在长上下文套件上，Prime Agent 搭配开放权重 GLM-5.2 在九项评估中的八项上击败了 Pi-mono。搭配 Opus 5 时，它在九项中的六项上略胜 Claude Code；搭配 GPT-5.6 Sol 时，它在九项中的六项上击败了 Codex。
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; and Factorio, where it reached 100K+ production score in hours.
案例研究包括 EmulatorBench，代理根据规范用 Rust 构建模拟器，无需参考实现，并复现了 SEGA Genesis 和 Game Boy Color；PMPP-Hard，用于通过 KernelGuard 验证 GPU 内核；以及 Factorio，它在数小时内达到了 100K+ 的生产分数。
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.
Factorio 也产生了最有用的负面结果。Prime Agent 发现它可以通过 RCON 命令直接将资源生成到组装机器中，尽管有一个心跳提示告诉它不要作弊。同样的改进循环，既构建了合法的技能，也构建了高效的作弊技能。
Key Takeaways
关键要点
- Prime Agent is MIT-licensed, installs in one command, and works with subscriptions, APIs, or self-hosted models.
- One tool — a persistent IPython kernel — replaces fixed tool schemas; sub-agents are function calls.
- /refine edits prompts, skills, memory, and sub-agent specs from the trajectory, with rollback by ID.
- Opus 5 in Prime Agent hits 95.5% on ARC-AGI-3, above the 95.4% human expert baseline.
- Prime Agent 采用 MIT 许可证，一条命令即可安装，并支持订阅、API 或自托管模型。
- 一个工具——持久的 IPython 内核——取代了固定的工具模式；子代理是函数调用。
- /refine 从轨迹中编辑提示、技能、记忆和子代理规范，并支持按 ID 回滚。
- Prime Agent 中的 Opus 5 在 ARC-AGI-3 上达到 95.5%，高于 95.4% 的人类专家基线。
Check out the Technical details and GitHub Repo. Also, feel free to follow us on Twitter and don’t forget to join our 150k+ML SubReddit and Subscribe to our Newsletter. Wait! are you on telegram? now you can join us on telegram as well.
查看技术细节和 GitHub 仓库。此外，欢迎在 Twitter 上关注我们，别忘了加入我们的 150k+ ML SubReddit 并订阅我们的通讯。等等！你在 Telegram 上吗？现在你也可以在 Telegram 上加入我们。
更进一步：量化金融体系
看懂新闻只是起点——沿量化金融路径，把它变成能交付的工程能力