AI Daily Digest — July 12, 2026: GPT-5.6 Goes Public, Muse Spark 1.1 Arrives, Open Robotics Pipeline
AI Daily Digest — July 12, 2026: GPT-5.6 Goes Public, Muse Spark 1.1 Arrives, Open Robotics Pipeline
AI 每日摘要 — 2026年7月12日：GPT-5.6 正式发布，Muse Spark 1.1 登场，开源机器人管线
🤖💻 AI Daily Digest — July 12, 2026 Another packed week in AI. OpenAI ended its 12-day restricted preview and opened GPT-5.6 to the world — three models, a new durability concept, and the ChatGPT Work + Codex integration that signals where the company is headed. Meta’s Muse Spark 1.1 landed with enough firepower to pull Mark Zuckerberg back to X after three years of silence. And NVIDIA and Hugging Face took a big swing at open-source robotics. Let’s dig in. 🤖💻 AI 每日摘要 — 2026年7月12日。AI 领域又度过了充实的一周。OpenAI 结束了为期 12 天的受限预览，正式向全球开放 GPT-5.6——包含三款模型、全新的“持久性”概念，以及 ChatGPT Work 与 Codex 的集成，这些都预示了该公司的未来方向。Meta 的 Muse Spark 1.1 表现强劲，甚至让马克·扎克伯格在沉寂三年后重返 X 平台。此外，NVIDIA 和 Hugging Face 在开源机器人领域也迈出了重要一步。让我们深入了解。
1. OpenAI GPT-5.6 Goes Public — Sol, Terra, Luna Redefine the Tier System
1. OpenAI GPT-5.6 正式发布 — Sol、Terra、Luna 重新定义分级系统
OpenAI officially released the GPT-5.6 series on July 9, ending 12 days of restricted government preview. The three-model family — Sol, Terra, and Luna — introduces a new “durable capability tier” concept: the names identify capability levels, not versions, meaning Sol can be upgraded to a future GPT-5.7 while keeping its tier identity. OpenAI 于 7 月 9 日正式发布了 GPT-5.6 系列，结束了为期 12 天的政府受限预览。该系列包含三款模型——Sol、Terra 和 Luna，引入了全新的“持久能力分级”概念：名称代表能力等级而非版本号，这意味着 Sol 可以升级到未来的 GPT-5.7，同时保留其分级身份。
Sol, the flagship, sets new state-of-the-art across coding (80 on the Artificial Analysis Coding Agent Index, beating Claude Fable 5 by 2.8 points), cybersecurity (73.5% on ExploitBench vs GPT-5.5’s 47.9%), and knowledge work. It runs in three effort modes: default for cost efficiency, max for extended reasoning, and ultra which coordinates 4 parallel agents by default (scalable to 16). Pricing runs $5/$30 per million input/output tokens for Sol, $2.50/$15 for Terra, and $1/$6 for Luna. 旗舰模型 Sol 在编程（在 Artificial Analysis 编程智能体指数中得分为 80，领先 Claude Fable 5 2.8 分）、网络安全（ExploitBench 得分 73.5%，而 GPT-5.5 为 47.9%）以及知识工作方面均创下了新的行业标杆。它提供三种运行模式：默认模式（兼顾成本效率）、最大模式（用于深度推理）以及超强模式（默认协调 4 个并行智能体，最高可扩展至 16 个）。定价方面，Sol 为每百万输入/输出 Token 5 美元/30 美元，Terra 为 2.50 美元/15 美元，Luna 为 1 美元/6 美元。
Alongside the model launch, OpenAI merged Codex into the ChatGPT desktop app and introduced ChatGPT Work, a unified interface for chat, coding, and long-running agent tasks. A new Programmatic Tool Calling feature in the Responses API lets GPT-5.6 write and run lightweight programs that coordinate tools inline. According to internal benchmarks, GPT-5.6 Sol improved the RSI Index by 16.2 points over GPT-5.5 on AI research acceleration tasks. 伴随模型发布，OpenAI 将 Codex 集成到了 ChatGPT 桌面应用中，并推出了 ChatGPT Work，这是一个集聊天、编程和长期智能体任务于一体的统一界面。Responses API 中新增的“程序化工具调用”功能，允许 GPT-5.6 编写并运行轻量级程序，从而在对话中直接协调工具。根据内部基准测试，GPT-5.6 Sol 在 AI 研究加速任务中的 RSI 指数较 GPT-5.5 提升了 16.2 分。
2. OpenAI GPT-Live — Real-Time Voice That Actually Listens and Speaks Simultaneously
2. OpenAI GPT-Live — 真正能同时听与说的实时语音模型
The same week, OpenAI launched GPT-Live, a full-duplex voice model that listens and speaks at the same time. Two versions — GPT-Live-1 and GPT-Live-1 mini — started rolling out globally on July 8. Previous voice systems either chained three models together (cascaded) or worked in rigid turn-based mode where the model waited for silence before responding. 同一周，OpenAI 推出了 GPT-Live，这是一款能够同时进行听与说的全双工语音模型。两个版本——GPT-Live-1 和 GPT-Live-1 mini——已于 7 月 8 日在全球范围内开始推送。此前的语音系统要么通过串联三个模型（级联），要么采用僵化的轮流对话模式，即模型必须等待用户停止说话后才能响应。
GPT-Live’s full-duplex architecture processes input continuously while generating output, making interaction decisions many times per second — whether to speak, listen, pause, interrupt, or invoke a tool. It handles backchannel cues (“mhmm”, “yeah”), stays quiet when you need a moment, and can perform real-time simultaneous translation. When a question requires deeper reasoning or search, GPT-Live delegates to GPT-5.5 behind the scenes and brings results back into the conversation without breaking flow. GPT-Live 的全双工架构在生成输出的同时持续处理输入，每秒做出多次交互决策——无论是说话、倾听、暂停、打断还是调用工具。它能处理反馈信号（如“嗯”、“是的”），在你需要思考时保持安静，并能进行实时同声传译。当问题需要更深层的推理或搜索时，GPT-Live 会在后台委派给 GPT-5.5，并将结果带回对话中，而不会中断交流流程。
3. Meta Muse Spark 1.1 — Zuckerberg Returns to X, Model Gets Serious About Agents
3. Meta Muse Spark 1.1 — 扎克伯格重返 X，模型全面发力智能体
Meta released Muse Spark 1.1 on July 9, and the event was significant enough that CEO Mark Zuckerberg posted on X for the first time in three years. “An incredibly capable agent and coding model at a very low price,” he wrote. Muse Spark 1.1 is purpose-built for agentic tasks — planning, tool calling, subagent delegation, and computer use. Meta 于 7 月 9 日发布了 Muse Spark 1.1，此次发布意义重大，以至于 CEO 马克·扎克伯格三年来首次在 X 上发帖。他写道：“这是一个能力极强且价格极低的智能体与编程模型。”Muse Spark 1.1 专为智能体任务而设计，包括规划、工具调用、子智能体委派和计算机操作。
On agent benchmarks, it scores 54.7% on JobBench (beating Claude Opus 4.8’s 48.4%) and 88.1 on MCP Atlas (ahead of Opus 4.8’s 82.2). Its 1-million-token context window and context compaction mechanism allow it to maintain state across long sessions. The model supports a main-agent/sub-agent delegation pattern, zero-shot generalization to new tools and MCP servers, and three computer-use execution modes (scripts, clicks, or batched actions per step). 在智能体基准测试中，它在 JobBench 上得分为 54.7%（超过 Claude Opus 4.8 的 48.4%），在 MCP Atlas 上得分为 88.1（领先 Opus 4.8 的 82.2）。其 100 万 Token 的上下文窗口和上下文压缩机制使其能够在长会话中保持状态。该模型支持主/子智能体委派模式、对新工具和 MCP 服务器的零样本泛化，以及三种计算机操作执行模式（脚本、点击或每步批量操作）。
4. Microsoft Swaps In-House MAI Models Into Excel and Outlook
4. 微软在 Excel 和 Outlook 中启用自研 MAI 模型
Microsoft has quietly started replacing third-party AI models — including OpenAI’s and Anthropic’s — with its in-house MAI series in core Office products, Bloomberg reported on July 8. Excel and Outlook now process tens of thousands of weekly AI prompts entirely on MAI models, a deployment scale not previously disclosed. 据彭博社 7 月 8 日报道，微软已悄然开始在其核心 Office 产品中用自研的 MAI 系列模型替换第三方 AI 模型（包括 OpenAI 和 Anthropic 的模型）。目前，Excel 和 Outlook 每周处理的数万条 AI 提示词已完全由 MAI 模型完成，这一部署规模此前并未披露。
5. NVIDIA and Hugging Face Open Up Humanoid Robotics
5. NVIDIA 与 Hugging Face 联手推动人形机器人开源
NVIDIA and Hugging Face announced on July 7 a major expansion of their robotics partnership, integrating NVIDIA’s Isaac GR00T 1.7 vision-language-action model and Isaac Teleop data-capture framework into Hugging Face’s open-source LeRobot library. GR00T 1.7 is the first open, commercially viable robot foundation model for humanoid robots. NVIDIA 和 Hugging Face 于 7 月 7 日宣布扩大双方在机器人领域的合作，将 NVIDIA 的 Isaac GR00T 1.7 视觉-语言-动作模型及 Isaac Teleop 数据采集框架集成到 Hugging Face 的开源 LeRobot 库中。GR00T 1.7 是首个面向人形机器人的开放且具备商业可行性的机器人基础模型。