最新资讯
techcrunch.com Airbnb 称将于 2027 年上半年推出 AI 旅行代理 theverge.com Perplexity 推出面向工作的 AI 代理 Teammates unite.ai Five9 拟以约3.25亿美元收购 Level AI,扩展客服AI分销渠道 unite.ai Serve Robotics 向物理AI公司开放车队软件与数据 wired.com 聊天机器人广告时代到来:哪些做法有效、哪些无效 techmeme.com METR:Anthropic 删减了评估报告中最具信息量的证据 unite.ai 微软借助 AI 修复更多 Windows 漏洞,但补丁部署仍滞后 bloomberg.com Cloudflare 上调年度利润预期,AI 推动网络需求 theinformation.com Anthropic 为 Claude 推出个人理财功能 theinformation.com Airbnb 披露 Chesky 隐秘 AI 初创公司的新细节 engadget.com Google 将停用 Firebase Studio,引导开发者转向 Antigravity 与 AI Studio www.youtube.com 别再给 AI Agent 发 API Key:Keycard 与 Temporal 谈 Agent 安全与可恢复执行
多元动态
登录 注册
AI编码 www.aibase.com ·

MiniMax 开源 Code CLI:命令行编程工具公布 76.7% 任务通过率

原文 我的收藏

MiniMax 正式开源其命令行 AI 编程工具 MiniMax Code CLI v0.4.12,源码已发布在 GitHub。官方称该工具在 FrontierHarness Eval 基准中任务成功率为 76.7%,成功任务中位耗时 4 分 33 秒,均优于公开基线。开源目的是让开发者检查调用逻辑与权限处理,共同构建更安全可靠的企业级 harness。

AI推荐理由 MiniMax 开源 Code CLI v0.4.12,源码上 GitHub,并给出 FrontierHarness Eval 76.7% 任务成功率和成功任务中位耗时 4 分 33 秒,属于可核验的官方开源动作与基准数据;但正文来自 AIbase 转述,未提供 GitHub 链接、基准细节或第三方复现,76.7% 与 4 分 33 秒只能按官方口径理解,不能视为独立验证。对关注编程 Agent、CLI 工具链和权限/调用透明度的开发者有实际参考价值。

72

正文

MiniMax Open Source Code CLI: Command-Line Programming Tool Achieves 76.7% Success Rate, Laying Enterprise-Level Harness in the Sunlight AIbase基地 Published inAI News · 3 min read · Sep 20, 2026 7 MiniMax has officially open-sourced the v0.4.12 version of its command-line AI programming tool, MiniMax Code CLI, with the source code now available on GitHub. As a core component of the MiniMax Code client, it is no longer just a product refined behind closed doors but now directly hands over the underlying implementation to developers. Its performance speaks for itself. In the FrontierHarness Eval benchmark, the CLI achieved a task success rate of 76.7%, with the median time for successful tasks reduced to 4 minutes and 33 seconds. Both metrics outperform the public baseline—meaning it not only works correctly but also runs efficiently. The official explanation for the open-source initiative is straightforward: it allows developers to personally examine the tool's call logic and permission handling, helping build a more secure and reliable enterprise-level harness together. For an intelligent agent that needs to write code and manipulate files in real production environments, the boundaries of permissions and the transparency of calls are precisely the aspects users cannot afford to overlook. By opening up the code, it is both a way to establish trust externally and to turn security reviews into a community-driven effort. With this move, MiniMax has effectively lifted the hood on its own programming intelligent agent's "cockpit." MiniMaxCodeCLIAIprogrammingtoolopensourceFrontierHarnessEval This article is from AIbase Daily Scan to view Welcome to the [AI Daily] column! This is your daily guide to exploring the world of artificial intelligence. Every day, we present you with hot topics in the AI field, focusing on developers, helping you understand technical trends, and learning about innovative AI product applications. —— Created by the AIbase Daily Team © Copyright AIbase Base 2024, Click to View Source -

来源报道 · 1 篇 · 以下链接均前往原始媒体

← 返回精选动态 原文

更多内容