苹果发布 HomeView 智能家居设备，配备 7 英寸屏幕和底座扬声器，售价 299 美元，即日起开启预购，11 月 13 日上市。该设备让升级版 Siri 进入家庭场景，可结合短信和日历上下文显示信息，并能通过人脸识别最多六名家庭成员，对话内容可在多设备间同步。苹果称设备摄像头和麦克风的数据与系统其余部分分开处理，用户被识别后才会在设备上显示其消息或日历内容。
#产品更新#语音#端侧
Catch it at the first user, not the four-hundredth. A new user’s bank link fails, the agent says “try again later,” and nothing shows up in the logs. Watch a team find the bug and fix it before anyone else hits it, with Agent Learning Services and Agent Data Services, launching today.
#产品更新#Agent#部署/工程
We've unified plans across Grok, Cursor, and X Premium. Introducing The X Subscription. https://t.co/SvkrVvBa0p
#行业动态#OpenAI#Anthropic
We're open sourcing agentchat, it's a unified SDK that allows you to connect your agent to Slack, Teams, iMessage, email, and WhatsApp in 2 lines of code and works with any agent SDK We noticed we kept on rebuilding the stack to connect our agents to different channels and ended up building the same shared layer. That's where agentchat came in. This allows you as a developer to focus on your core agent's brain and let the SDK manage your connections with different channels.
#开源/仓库#Agent#MCP/工具调用
Jev as a judge for LLM evals! JevEval uses Jev as the judge for custom LLM evals instead of relying on another generative LLM for every evaluation. The idea is pretty simple. You define the exact decisions you want the evaluator to make, Jev evaluates each one and returns probabilities, and DeepEval uses those results to calculate the final score. For example, with an AI agent you can evaluate the trace itself, checking whether the agent selected the right tools, followed the expected sequence of actions, and actually completed the task. Jev handles those semantic judgments, while the scoring logic stays outside the model. That is the main difference from a normal LLM-as-a-judge setup. With a generative judge, the model often has to interpret the rubric, reason about the response, and generate the evaluation itself. With JevEval, you define the decision structure upfront. You can also assign different weights to different checks or use `strict_mode` when every requirement needs to pass. So if one part of the agent workflow matters more than another, you can reflect that directly in the scoring logic instead of leaving the entire evaluation to the judge model. Key capabilities: - Evaluate agent outputs and traces with Jev - Define custom checks as bounded decisions - Use yes/no, score, and closed-choice questions - Weight different checks or require all of them to pass - Evaluate both single-turn outputs and multi-turn conversations This is where Jev makes a lot of sense for evals. If you already know what needs to be checked, you can define those decisions upfront, let Jev handle the semantic judgment, and keep the final scoring deterministic. I’ve shared the DeepEval post in the comments.
#评测/基准#Agent#产品更新
#行业动态#OpenAI#xAI
You already gave the LLM the answer. So why is it writing it out from scratch? Introducing a new AI model architecture: Constrained Decoding Models (CDMs) LLMs were built to write. Most production workloads just need an answer: fast, consistent, and affordable. CDMs are purpose-built for 1 task each. They read your input once and skip the token-by-token generative loop that makes LLMs slow and expensive. Not just a decision model, but a family of 3 CDMs live today: • Classify: define your labels and get back a calibrated probability + reason for each. Used for routing, tagging, and intent detection. • Extract: describe the entities in natural language and get back well-formed JSON with confidence scores + citations. Used for web scraping, document processing, and PII redaction. • Summarize: pass in long text, images, or PDFs and get back a consistent summary under 1,000 tokens. Used for context compaction, metadata generation, and transcript digests. Built for production: • 1M tokens of multimodal (text + image) input, 31x larger than Jev • 77.2% on DecisionBench vs Jev's 72% • 200x faster, 467x cheaper than frontier LLM APIs • $0.05/M input tokens. Output is free (yes, even for summarize) Live on your favorite platforms today: @OpenRouter @vercel_dev @LiteLLM @LangChain @TrustedRouter @itsArthurAI @useRouterPlus Get 50M input tokens for free (for the first 50 hours) at link below 👀
#产品更新#部署/工程#OpenAI
Introducing SubQ Snap, the first long-context decision model for real enterprise work. Snap is the only decision model or LLM where compute per token stays constant as context grows. • 384K tokens of multimodal context, 12x larger than Jev. • 61.7 on Decision Index v0.2.1 vs Jev’s 57.9, outperforming on 13 of 16 enterprise benchmarks. •Processes 100K tokens and returns a decision in under one second. •$0.05/MM input tokens. Output is free. Available today to anyone. The first 10,000 sign-ups get 100M free input tokens. We’ll likely hit capacity fast and will add users as we add compute. Compute offers welcome 🙂. Latency will be variable as we tune the platform to this new model and category of users. Model speeds are 80ms for minimal context, 250ms for 32k, 660ms for 100k, and 2.5 seconds for 384k. Visit: https://t.co/Nc76iJF2a8
#模型发布#推理#多模态
We've unified plans across Grok, Cursor, and X Premium. Introducing The X Subscription. https://t.co/SvkrVvBa0p
#行业动态#OpenAI#产品更新
We've unified plans across Grok, Cursor, and X Premium. Introducing The X Subscription. https://t.co/SvkrVvBa0p
#行业动态#编码#OpenAI
You already gave the LLM the answer. So why is it writing it out from scratch? Introducing a new AI model architecture: Constrained Decoding Models (CDMs) LLMs were built to write. Most production workloads just need an answer: fast, consistent, and affordable. CDMs are purpose-built for 1 task each. They read your input once and skip the token-by-token generative loop that makes LLMs slow and expensive. Not just a decision model, but a family of 3 CDMs live today: • Classify: define your labels and get back a calibrated probability + reason for each. Used for routing, tagging, and intent detection. • Extract: describe the entities in natural language and get back well-formed JSON with confidence scores + citations. Used for web scraping, document processing, and PII redaction. • Summarize: pass in long text, images, or PDFs and get back a consistent summary under 1,000 tokens. Used for context compaction, metadata generation, and transcript digests. Built for production: • 1M tokens of multimodal (text + image) input, 31x larger than Jev • 77.2% on DecisionBench vs Jev's 72% • 200x faster, 467x cheaper than frontier LLM APIs • $0.05/M input tokens. Output is free (yes, even for summarize) Live on your favorite platforms today: @OpenRouter @vercel_dev @LiteLLM @LangChain @TrustedRouter @itsArthurAI @useRouterPlus Get 50M input tokens for free (for the first 50 hours) at link below 👀
#产品更新#Agent#部署/工程
#行业动态#端侧#部署/工程
#产品更新#Agent#部署/工程
腾讯混元发布科学发现智能体 Hyra，采用 Context Agent 维护经验库、多个 Proposal Agent 生成并沙盒评估 solution 的异步流水线，无评估器时升级为双层循环让评估器与 solution 共进化。在 NanoChat Autoresearch、NanoGPT Speedrun 和 SOL-ExecBench 三项任务上，Hyra-1.0 分别取得 Validation BPB 0.9015、Time to 3.28 loss 76.4s 和 Mean SOL 0.771，均超过 Recursive 报告的结果。在 55 个数学开放问题中，Hyra 于 29 个上刷新历史最好结果，相关结果已在 GitHub 开源。
#模型发布#Agent#推理
InternLM 开源 InternBootcamp（InternAgentHarness），一个面向大语言模型智能体训练与评测的可扩展合成环境框架，将任务提示生成、工具执行、交互控制和奖励计算统一封装为可训练的环境实例。框架提供 BootCampCLI，可把任务配置转换为可运行的智能体交互流程，并支持批量轨迹采集、监督微调数据过滤、强化学习 rollout 和测试集评测。其技术报告已被 WAICA 2026 接收，对应代码版本为 InternBootcamp v1。
#开源/仓库#Agent#数据/训练
InternLM 发布开源多模态决策模型 Intern-Decision，将状态、可选图像和多个带类型的问题转化为带概率的决策，支持 choice、score 和 noul（是/否）三类问题，训练时微调 Qwen3.5 的语言主干并冻结视觉塔和投影器。Intern-Decision-4B 在七套基准上平均准确率 90.02%，RTX 4090 本地平均请求延迟 44.16 ms，在 96 例校准基准上 Brier 为 0.550、ECE 为 0.089。该版本包含训练、两种推理后端、基准评分、温度校准和浏览器 demo，但训练数据、校准与验证记录、图像及模型权重未包含在内。
#开源/仓库#多模态#推理
World Labs 分享了其早期进展，目标是生成持久、可导航的 3D 世界，并可直接在浏览器中探索。该内容发布于 2024 年 12 月 2 日，页面同时列出 2026 年 9 月 28 日的后续文章《To Seek a Newer World》，讲述从 World Labs 到 AMD 的历程。
#行业动态#多模态#产品更新
World Labs 推出 Marble，这是其 3D 世界生成模型的有限访问 beta 预览，用户可在 marble.worldlabs.ai 查看和创建 3D 世界。给定图像或文本提示，模型可生成可无限探索、无变形、无不一致的 3D 世界，相比此前结果更大、风格更多样、3D 几何更干净。用户可将生成的世界导出为 Gaussian splats，并借助开源渲染库 Spark 集成到 Three.js 中，在桌面、笔记本、移动设备和 VR 头显上渲染。
#产品更新#多模态#图像生成
World Labs 发布 World API，开发者可通过该公开接口调用其多模态世界模型 Marble，从文本、单图、多图、360° 全景和视频生成可导航的 3D 世界。世界生成以异步方式运行，无需专有工具，生成结果可在浏览器渲染、导出到其他工具或接入交互系统与仿真管线。该 API 已用于游戏与沉浸式媒体、机器人仿真（可接入 NVIDIA Isaac Sim、MuJoCo、RoboSuite）以及建筑设计等场景，并通过 World Labs Platform 提供 API key 管理、用量计费和文档。
#产品更新#多模态#部署/工程
World Labs 发布 Spark 2.0，为基于 THREE.js 和 WebGL2 的 3D Gaussian Splatting 渲染器加入 Level-of-Detail 系统，可在任意设备上流式渲染超大 3DGS 世界。新版本通过 LoD splat 树、渐进式流式加载和虚拟显存三项技术，把每帧渲染量控制在 500K 至 2.5M splats 的预算内，并引入新的 .RAD 流式文件格式。Spark 为开源项目，配套命令行工具 build-lod 可将任意 3DGS 文件转换为 LoD splat 树。
#产品更新#部署/工程#开源生态