Illustration depicting Anthropic and OpenAI launching AI agent teams amid a $285B software stock drop.
AI 生成的图像

Anthropic and OpenAI release AI agent management tools

AI 生成的图像

On February 5, 2026, Anthropic and OpenAI simultaneously launched products shifting users from chatting with AI to managing teams of AI agents. Anthropic introduced Claude Opus 4.6 with agent teams for developers, while OpenAI unveiled Frontier and GPT-5.3-Codex for enterprise workflows. These releases coincide with a $285 billion drop in software stocks amid fears of AI disrupting traditional SaaS vendors.

The launches mark a broader industry trend toward AI as a delegated workforce rather than a conversational partner. Anthropic's Claude Opus 4.6, succeeding the November 2024 Opus 4.5, features a 1 million-token context window in beta and excels on benchmarks like Terminal-Bench 2.0, ARC AGI 2 (68.8%), and MRCR v2 (76% at 1 million tokens). It outperforms OpenAI's GPT-5.2 and Google's Gemini 3 Pro in several evaluations, though OpenAI's new GPT-5.3-Codex scored 77.3% on Terminal-Bench 2.0. Pricing remains $5 per million input tokens and $25 per million output tokens via API.

Anthropic's 'agent teams' in Claude Code allow developers to create multiple AI agents that split tasks, coordinate autonomously, and run in parallel, ideal for read-heavy work like codebase reviews. Available as a research preview, it presents a split-screen interface for monitoring and intervening.

OpenAI's Frontier platform positions AI agents as 'co-workers' with individual identities, permissions, and memory, integrating with CRMs, ticketing tools, and data warehouses. Barret Zoph, OpenAI's B2B general manager, stated, 'What we’re fundamentally doing is basically transitioning agents into true AI co-workers.' Released alongside a macOS app for Codex—described as a 'command center for agents'—GPT-5.3-Codex supports the full software lifecycle, including debugging, deploying, and monitoring. It runs 25% faster and was 'instrumental' in its own development, such as managing deployments and test results.

These tools require human oversight to correct errors, and no independent tests confirm they outperform solo developers. The timing follows Anthropic's January 30 release of 11 open-source plugins for its Cowork tool, which extend Claude into domains like legal review and financial analysis. Investors reacted with panic, erasing $285 billion from software, financial services, and asset management stocks; a Goldman Sachs software basket fell 6%. OpenAI's Fidji Simo emphasized, 'Frontier is really a recognition that we’re not going to build everything ourselves.' Anthropic's Scott White dubbed the shift 'vibe working.'

Opus 4.6 is accessible on Pro ($20/month), Max, Team, and Enterprise plans, with usage limits.

人们在说什么

X discussions celebrate Anthropic's Claude Opus 4.6 agent teams and OpenAI's simultaneous GPT-5.3-Codex and Frontier launches as a paradigm shift to managing AI coworkers over chatting. Developers highlight parallel execution, 1M token context, and autonomous coding gains. Investors note fears of SaaS extinction amid $285B software stock plunge. Skeptics debate data moats, legal risks, and true replacement of traditional tools.

相关文章

Professional using Anthropic's Claude AI Cowork feature on MacBook to automatically organize files and generate reports, as shown in a realistic office scene.
AI 生成的图像

Anthropic launches Cowork feature for Claude AI

由 AI 报道 AI 生成的图像

Anthropic has introduced Cowork, a new tool that extends its Claude AI to handle general office tasks by accessing user folders on Mac computers. Designed for non-developers, it allows plain-language instructions to organize files, create reports, and more. The feature is available as a research preview for Claude Max subscribers.

Anthropic has announced that its AI chatbot Claude will remain free of advertisements, contrasting sharply with rival OpenAI's recent decision to test ads in ChatGPT. The company launched a Super Bowl ad campaign mocking AI assistants that interrupt conversations with product pitches. This move highlights growing tensions in the competitive AI landscape.

由 AI 报道

2025 年,AI 代理成为人工智能进步的核心,使系统能够使用工具并自主行动。从理论到日常应用,它们改变了人类与大型语言模型的互动。然而,它们也带来了安全风险和监管空白等挑战。

OpenClaw, an open-source AI project formerly known as Moltbot and Clawdbot, has surged to over 100,000 GitHub stars in less than a week. This execution engine enables AI agents to perform actions like sending emails and managing calendars on users' behalf within chat interfaces. Its rise highlights potential to simplify crypto usability while raising security concerns.

由 AI 报道

Google has released Gemini 3 Pro, its latest flagship AI model, emphasizing improved reasoning, visual outputs, and coding capabilities. The company also introduced Antigravity, an AI-first integrated development environment. Both are available in limited preview starting today.

In a comparative evaluation of leading AI models, Google's Gemini 3.2 Fast demonstrated strengths in factual accuracy over OpenAI's ChatGPT 5.2, particularly in informational tasks. The tests, prompted by Apple's partnership with Google to enhance Siri, highlight evolving capabilities in generative AI since 2023. While results were close, Gemini avoided significant errors that undermined ChatGPT's reliability.

由 AI 报道

随着AI平台转向基于广告的变现模式,研究人员警告这项技术可能以隐形方式塑造用户行为、信念和选择。这标志着OpenAI的转变,其CEO Sam Altman曾认为广告与AI的结合“令人不安”,但现在保证AI应用中的广告能够维持信任。

 

 

 

此网站使用 cookie

我们使用 cookie 进行分析以改进我们的网站。阅读我们的 隐私政策 以获取更多信息。
拒绝