https://ai-tokens.cn/news/news-meta-muse-20260922.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 Meta 个人 AI 智能体 Muse 上线两周登顶 App Store，单日市值飙升近 2000 亿美元 https://ai-tokens.cn/news/news-gpt-6-astra-20260903.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 OpenAI 发布 GPT-6 Astra：「欢迎来到 AGI 时代」，API 定价 $10/$50 为上代 2.5 倍 https://ai-tokens.cn/news/news-gemini-38-flash-20260903.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 Google 三周内再推 Gemini 3.8 Flash：介绍价 $0.75/$3.75，Cyber 网络安全变体齐发 https://ai-tokens.cn/news/news-workbuddy-open-platform-20260903.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 WorkBuddy 开放平台解析：用 Skill、Connector、专家与知识库给 AI 同事「配工具」 https://ai-tokens.cn/news/news-wechatpay-ai-card-20260831.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 微信支付 AI 专属卡上新：一张卡直连 DeepSeek Harness 与 OpenClaw https://ai-tokens.cn/news/news-anthropic-mhs-20260828.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 Anthropic 发布 MHS 模型硬件标准：让 AI 像插 USB-C 一样操控真实世界设备 https://ai-tokens.cn/news/news-dd2f144e562f02fa.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 清华特奖得主AI创业再+1：GeoPT一作入局物理世界模型 https://ai-tokens.cn/news/news-1bf5025870127a47.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 说真话就能拿10万？网易叭哥说重金求“真需求”，不卡专业学历 https://ai-tokens.cn/news/news-8ffbbf24c13a812d.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 India becomes the latest country to report rogue AI agent activity https://ai-tokens.cn/news/news-68b9f741f3df646e.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 An Nvidia insider-turned-VC on why robotics hasn’t hit its ChatGPT moment, yet https://ai-tokens.cn/news/news-bd2b33f50b82d969.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 Physical AI is more than robots, says the Nvidia veteran now investing in it https://ai-tokens.cn/news/news-5cea1aeddd4affb9.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 Elorian’s first AI model wants to reason with images, not just about them https://ai-tokens.cn/news/news-7ff2fab56ea772a2.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 Jev is Vercel’s fastest-adopted AI model ever https://ai-tokens.cn/news/news-e1b3c44f13458aa6.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 Mark Wahlberg on using AI in film ‘the right way’ and being open to an AI clone carrying on his legacy https://ai-tokens.cn/news/news-54ebdfe79863725a.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 A week with RTX Spark: everything we’ve learned using Nvidia’s new laptops https://ai-tokens.cn/news/news-8655f4cf26df3e21.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 ChatGPT becomes a full Chrome browser assistant that reads and acts on your open tabs https://ai-tokens.cn/news/news-33f25b1f0c847371.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 China plans to tighten export controls on open-weight models and 23 key AI technologies https://ai-tokens.cn/news/news-8d23fe4bf5d07911.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 Google's Gemini 4 Argon is closer to the AI frontier than any Google model in months https://ai-tokens.cn/news/news-4c64677b0f2c91e0.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 Veritas expands its anti-counterfeit chip to jewelry, trophies, and fragrance bottles https://ai-tokens.cn/news/news-ca580b1475b7847a.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 Meta’s new terms may allow it to train AI on news sites for free https://ai-tokens.cn/news/news-80ab6ac91ca5d150.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 OpenAI's new Rosalind Academic Plan gives research labs up to 30 free seats for a year https://ai-tokens.cn/news/news-052f9ed9e889d39c.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 Anthropic gives METR inside access to audit its AI models and safety practices https://ai-tokens.cn/news/news-afde317889b48fa7.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 Meta Releases Muse Spark Open: 816B Open-Weight MoE With 1M Context https://ai-tokens.cn/news/news-a828bd7f7dff23d5.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 The World Won't Wait: Evaluating and Teaching Agents Under a Live Clock https://ai-tokens.cn/news/news-64ab9f94b7d65ce3.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 Evaluating Model Capabilities After Benchmark Saturation https://ai-tokens.cn/news/news-b8fdb77fd6619c1e.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 Value-Informed Schedule Trimming for Accelerated Sequential Monte Carlo Guidance in Discrete Diffusion https://ai-tokens.cn/news/news-77ee249104a767fa.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 AlphaDiana: A Measurement System for Agents in Verifiable Reasoning Tasks https://ai-tokens.cn/news/news-a61630d2e5679d80.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 Counterfactual Neighborhood Risk Fields: Evidence-Grounded Selective Safety for Vision-Language Models https://ai-tokens.cn/news/news-5d6ab34c3d75d540.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 IneqAgent: Agentic Inequality Proving in Lean with Symbolic Computation https://ai-tokens.cn/news/news-91832747d64f6512.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 L2G-SR: Learning to Guide LLM-based Symbolic Regression https://ai-tokens.cn/news/news-c37f1987521f7498.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 When Reality Outpaces Cognition: A New Scaling Principle for Reliable Intelligence https://ai-tokens.cn/news/news-cffdb744dda380b1.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 ARIA: Runtime Inspection and Amendment for Multi-Agent Systems https://ai-tokens.cn/news/news-94d5d577ef0fe3a4.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 Open-world AI: a Tale of Three Requirements https://ai-tokens.cn/news/news-4faf84fcc2c0dce6.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 EvolvingBrain: An Adaptive Skill System for Multi-modal Agents Self-Evolution in Open Worlds https://ai-tokens.cn/news/news-94f07ba05cddef93.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 When Reasoning Traces Conflict with Evidence: How Answer Configuration Shapes Evidence Following https://ai-tokens.cn/news/news-90ec246428241617.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 A Self-Improving Agent that Grows its Own DSL: Oracle-Free Operator Induction for Exact Question Answering https://ai-tokens.cn/news/news-ae2eb56749d9d450.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 Temporal Specification Guidance for Text-to-Video Generation https://ai-tokens.cn/news/news-b91f8333bb510845.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 DeepUnfolding_Agent: A Classical_Quantum Adaptive Control API https://ai-tokens.cn/news/news-9f810e2e7868773b.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 PRISM: PRIvileged Sibling-guided Multi-Agent Topology Design https://ai-tokens.cn/news/news-eb9262ac4eece3cf.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 A Large-Scale Dataset for Multi-Class Financial News Classification in Brazilian Portuguese https://ai-tokens.cn/news/news-bb280405ef39d55c.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 MeddeX: A Retrieval-Augmented LLaMA-3 System with Gemini Vision for Evidence-Backed Medical Report Analysis https://ai-tokens.cn/news/news-8eed4925d1c0d15a.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 Jev-as-a-Judge for RAG Claim Verification: Confidence-Gated Cascading Between a Non-Generative Verifier and a Frontier LLM https://ai-tokens.cn/news/news-27faadd7cc1b81ba.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 Exploring the Design Space of Permutation Diffusion LLMs https://ai-tokens.cn/news/news-2e1d47a1b495434b.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 AbridgeKV: Compressing Long Reasoning into the KV Cache of Learned Pause Tokens https://ai-tokens.cn/news/news-e09eba81dad6d91a.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 CoTS: Competitive Token Selection for Latent Reasoning in Large Language Models https://ai-tokens.cn/news/news-591f825a8daf7206.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 Does Personalization Break Trust? Open-Set Rejection Calibration Under Few-Shot On-Device Intent Adaptation https://ai-tokens.cn/news/news-5dbdb451b813fb62.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 Reference-Guided Turn-Taking Style Transfer for Full-Duplex Spoken Language Models https://ai-tokens.cn/news/news-be3c075f36fe1970.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 Verdict: Encode Once, Decide Many https://ai-tokens.cn/news/news-5a1cf1b87684bcb3.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 APRiL: An Open-ended Benchmark for Chinese Social Awareness in Large Language Models https://ai-tokens.cn/news/news-5016d48969650b81.html AI导航 zh-CN 2026-10-17T02:56:31+08:00 PRISM: Transfer and Limits of Activation Interventions in Tool-Calling Language Models