Tag
LlamaIndex promotes LlamaParse's ability to cleanly extract nested tables (e.g., from Micron's earnings deck) into markdown, highlighting features like layout awareness, multimodal parsing, multilingual support, and enterprise readiness.
Asana 通过在 Codex 中使用 GPT-6 Astra 优化其 StackAI 浏览器代理工作流，在 GPT-6.1 Sol 上实现了成本降低 76 倍、速度提升 5 倍，原本需要一到两个月的手动研究工作缩短至约一周。
该论文提出「先澄清、再聚焦」的声明归一化方法，将对话转化为带来源引用和语义标签的简短陈述，使仅由小模型构成的推理管线能以约100倍更低的成本处理数百万条企业客服对话的分析任务。
Google Cloud 在 Gemini at Work 2026 上发布 Gemini agent——一个面向工作场景的统一智能体，可在单一提示框中回答问题、处理知识工作、创建媒体、编写并运行代码，并深度集成 Google Workspace、行业专属技能与企业级安全治理能力。
Gemini Agent 是一款面向工作场景的通用 AI 代理，每个任务拥有独立的计算机环境，支持在 Gemini 与 Claude 模型之间自由选择，并具备子代理、多层记忆、主动式工作流、企业级治理与成本上限等功能。
Google 在 Google Cloud 活动上宣布为 Gemini 推出统一的企业级 agentic AI 助手，可自主规划任务、调用第三方模型（如 Anthropic 的 Claude）并连接企业内部系统，初期面向商业用户开放。
Google is launching a universal Gemini AI agent for enterprise users that can work across apps, devices, and third-party tools like Slack and Microsoft 365 from a single interface, currently in private preview.
MIT Technology Review's Business Lab examines how advances in foundation models, physical AI, and agentic AI are reshaping industrial automation, with AVEVA's Arti Garg discussing responsible deployment, human oversight, and an IEEE methodology for measuring AI's environmental footprint.
Nous Research confirmed a $90M Series B at a $1.5B valuation, led by Robot Ventures with backers including Nvidia and Samsung, and announced Hermes for Businesses to bring its open-source AI agents into enterprise workflows.
GitHub Copilot's local sandboxing feature is now generally available, letting Copilot run commands in an isolated environment with controlled access to files, networks, and credentials. Enterprise teams can centrally manage policies across the Copilot CLI, app, and VS Code.
Bret Taylor announces the Personal Agent Protocol, an open standard developed with Meta, Sierra and partners like Shopify, Stripe and Walmart to define how personal agents interact with businesses, while a commenter questions whether MCP with access control already covers that need.
LangChain announced a webinar on the agentic operating model, covering how enterprise teams should structure ownership, standardization, governance, and infrastructure for scaling production AI agents.
Anthropic expanded its Claude for Startups program, offering qualifying early-stage companies a free year of Claude Team with up to five premium seats and $1,000 in API credits, plus access to Claude Marketplace and virtual office hours with its Applied AI team.
OpenAI announced a partnership with Ironclad to turn real contracting and procurement workflows into training and evaluation tasks for AI agents, with GPT-6 Astra scoring 32% higher than GPT-5.6 Sol on these tasks while cutting estimated time per attempt by 48%.
UndoBench introduces a benchmark that decouples task competence from fault-recovery capability in tool-using AI agents across 36 workflows and 36 fault scenarios in 8 enterprise domains, finding nominal task completion of 83.54% masks conditional recovery success of just 46.72%, with naive retry producing duplicate side effects in 53.33% of trials.
eu/jev launches as the first System One model hosted in the EU, running on Hetzner GPUs in Germany with full GDPR compliance, no prompt logging, and pricing identical to Jev (including free output tokens). It builds on Winnow, billed as the top open-source System One model on JevBench, aimed at enterprises blocked by EU data-residency requirements.
Liquid AI's D1 decision model is now live on OpenRouter, offering typed probabilistic answers for state-based decision queries at $0.04/M input with zero data retention. The community is hyped to stress-test OpenRouter's serving infrastructure with the new model.
Microsoft is rethinking Copilot, repositioning it as the 'OS for work' with unified consumer/enterprise UI, coding and agent features, and the full Office suite, while consolidating leadership under Jacob Andreou and rebranding Microsoft Scout to Autopilot.
E2B introduces E2B Embed, an open-source (Apache-2.0) way to package the E2B Runtime and run Firecracker-based sandboxes inside a customer's own environment, aimed at regulated industries and government clients requiring data residency. It offers Docker Compose, Terraform (AWS/GCP), and Kubernetes deployment options with the same Python/JavaScript SDKs, a dashboard, and OpenTelemetry export.
Anthropic's Claude for Government is now generally available for federal and state agencies, delivering coding and agentic capabilities through a FedRAMP High authorized environment with usage-based billing, admin controls, and audit logging. Claude Code CLI and Claude for Microsoft 365 are also rolling out in early access through the same environment.