全部 AI 动态

Adaptive Workflow Intelligence: A Cognitive Architecture for Context-Driven Enterprise Automation

arXiv:2610.08793v1 Announce Type: new Abstract: Enterprise systems increasingly rely on automated workflows, yet many AI-driven solutions remain brittle under non-stationary conditions, evolving policies, and delayed operational feedback. While reinforcement learning and large language model (LLM) agents offer partial adaptability, they do not by themselves provide persistent reflection mechanisms or straightforward …

Accelerating Floating-Point Satisfiability Solving via Gradient Normalization

arXiv:2610.08808v1 Announce Type: new Abstract: Satisfiability Modulo Theories (SMT) solvers are foundational to software verification, program analysis, and compiler testing, particularly over the theory of Quantifier-Free Floating-Point (QF_FP). While recent optimization-based SMT solvers have successfully applied gradient descent to continuous relaxations of logical formulas, they are fundamentally bottlenecked by…

Route-Verify-Vote: Procedure-Conditioned Self-Consistency for Mixed-Domain Reasoning

arXiv:2610.08814v1 Announce Type: new Abstract: Compositional generalization remains challenging when language models must combine familiar reasoning operations in unfamiliar ways. The Scenario-Based Commonsense Reasoning Evaluation (SCoRE) 2026 tests this ability on three mixed domains absent from training and requires models to identify the complete set of correct options for each question. We introduce Route-Verif…

An Empirical Study of Agent Skills' Downstream Utility

arXiv:2610.08875v1 Announce Type: new Abstract: Agent Skills package procedural guidance and resources for reuse, but a relevant Skill does not necessarily improve task performance. Existing studies characterize Skill content and evaluate downstream performance, yet provide limited explanations of how utility depends on content, execution configuration, and multi-Skill organization. We conduct an empirical study on 8…

How Could AI Eliminate Humanity? A Failure-Mode Analysis of Civilizational Risk

arXiv:2610.08878v1 Announce Type: new Abstract: This article develops a failure-mode framework for analyzing how advanced artificial intelligence could contribute to human extinction, irreversible civilizational collapse, or permanent human disempowerment. The central thesis is that catastrophic AI risk does not require consciousness, hostility, or an explicit intention to harm humanity. Instead, risk may arise throu…

Humanize: Judgement Engineering for Agentic Coding

arXiv:2610.08900v1 Announce Type: new Abstract: Agentic coding makes code generation cheap, but reliable completion remains difficult: the agent that writes the code is a weak judge of whether it is done. We present Humanize, a multi-agent orchestration workflow for agentic coding built around judgement engineering: explicit, mechanically enforced decisions at the boundaries between planning, implementation, review, …

Sequential Probabilistic Uncertainty Estimation for Parallel Multi-Agent Reasoning Systems

arXiv:2610.08901v1 Announce Type: new Abstract: LLM-based multi-agent systems (MAS) have attracted growing attention for improving reasoning through interaction among multiple agents. In this work, we focus on parallel multi-agent reasoning systems, where several agents solve the same problem over multiple rounds and aggregate their outputs into a final answer. Despite their strong reasoning performance, uncertainty …

Agent Plasticity: Measuring Self-Improvement Through Experience

arXiv:2610.08902v1 Announce Type: new Abstract: AI agents increasingly operate in environments where they can diagnose failures and improve through experience, yet existing evaluations largely measure what an agent can do at a fixed point in time rather than how effectively it learns. Evaluating self-improvement requires answering three questions: does future performance improve and generalize beyond the interactions…

AdaGuard: Enhancing Safety and Policy Compliance with Reasoning-Enabled LLM-As-A-Judge Guardrails

arXiv:2610.08923v1 Announce Type: new Abstract: Enterprise generative AI applications require robust safety mechanisms that can accommodate diverse risk postures, evolving policies, and varying latency constraints. Current guardrail solutions often suffer from rigidity, relying on fixed policy sets and offering limited transparency or reasoning flexibility. We present Adaguard, an adaptive LLM-as-a-Judge framework de…

Can AI Agents Make Open-Ended Scientific Discovery? Evidence from Station

arXiv:2610.08927v1 Announce Type: new Abstract: Recent AI systems have made rapid progress in scientific discovery when given well-defined metrics, but whether they can autonomously undertake open-ended scientific discovery remains unclear. We investigate AI's ability to tackle open-ended tasks in Station, an open-world environment in which multiple agents simulate a scientific ecosystem. To tackle challenges specifi…

Humanity's Sixth Sense: Benchmarking Intuitive Visual Reasoning in Multimodal Models

arXiv:2610.08966v1 Announce Type: new Abstract: Humans perceive far more in a scene than what is explicitly depicted: a single glance captures past causes and future trajectories; a quick peek determines if a vehicle can fit between two parked cars; a few seconds of video reveals who holds authority in a room; and a fleeting clip highlights subtle abstract patterns like unwritten rules or hidden labels. This capacity…

Socio-Foundation: A Model for Generalizable Individual Behavior Simulation via Hierarchical Capability Distillation

arXiv:2610.08967v1 Announce Type: new Abstract: Simulating individual behavior requires large language models (LLMs) to preserve persona traits while adapting to dynamic social contexts. However, general-purpose LLMs often flatten distinct personas, while task-specific tuning suffers from fragmentation and generalization. To overcome these challenges, we organize individual simulation into the \textbf{FONTS Taxonomy}…

Verify Less, Evolve More: Training Idea-Level Critics for Verification-Efficient ML Evolving Agents

arXiv:2610.08993v1 Announce Type: new Abstract: As large language models become more powerful, self-evolving agents are able to tackle challenging tasks including AI for machine learning (AI4ML). In AI4ML, while empirical verification is available, it often requires computationally costly model training and evaluation, limiting the speed and scale of agent evolution. Yet verification efficiency remains under-explored…

How Fragile Is On-Device Language Model Safety? Localizing Safety-Critical Parameters for Sparse Fault Analysis

arXiv:2610.09000v1 Announce Type: new Abstract: As small language models (SLMs) are increasingly deployed on resource-constrained and on-device platforms, including as components of agentic systems, the integrity of locally stored model parameters becomes an important safety concern. We investigate whether safety-sensitive behavior in LLaMA-2-7B-Chat is concentrated within a sparse subset of parameters, creating a re…

Learning to Report Unsafe Tasks in a Multi-Agent Game

arXiv:2610.09002v1 Announce Type: new Abstract: When agents share a reward for completed tasks, reporting unsafe work can reduce the reporter's reward by stopping a task. Audits can make reporting optimal without ensuring that further training teaches a silent team to report. We study this learning problem in a game where any witness can stop a task by reporting. With $k$ witnesses per task sharing a policy and drawi…

Sigma-Hunter: A Domain-Specific Language Model for Threat Hunting and Detection Engineering

arXiv:2610.09007v1 Announce Type: new Abstract: Detection engineers must translate threat reports, forensic observations, and hunt hypotheses into precise, testable rules. General-purpose large language models (LLMs) can draft such rules, but often produce invalid YAML, incorrect log sources, unsupported fields, or overly broad detection logic. This paper presents \emph{Sigma-Hunter}, a domain-adapted LLM for analyst…

Whose Memory Is It? Scope-Aware Commit Rules for Long-Term LLM Memory

arXiv:2610.09008v1 Announce Type: new Abstract: Persistent memory allows an LLM agent to carry experience across conversations, but it also turns a local reasoning mistake into a durable one. During deliberation, an agent may consider a plan, simulate a tool result, report another speaker's belief, and then reject all of them. If memory retains only the resulting sentences, those once-useful possibilities can later r…

Enabling Dynamic Computation in Looped LMs

arXiv:2610.09013v1 Announce Type: new Abstract: Looped LMs are parameter efficient and promise dynamic computation (saving memory and FLOPs on easy tokens). However, state-of-the-art open Looped LMs trained with this dynamic computation capability (Ouro models) do not realize it in practice as each loop iteration (depth) requires its own level of KV-cache, necessitating all loop computations. Moreover, Ouro's early-e…

PAIR: Bridging Perception and Action in Vision-Language-Action Models

arXiv:2610.09016v1 Announce Type: new Abstract: Vision-language-action (VLA) models map visual observations and language instructions to continuous robot actions. This task requires a transition from representations that describe the scene and instruction to representations that support action generation. Many continuous-action VLAs leave this transition implicit and supervise it mainly through the final action-predi…

Not Every Call Needs a Frontier Model: Per-Call-Site Evaluation of Small Language Models in a Deployed Agentic Home-Automation System

arXiv:2610.09021v1 Announce Type: new Abstract: An agentic system issues several structurally different kinds of LLM calls. It routes intent, classifies actions, grounds language in a device registry, plans multi-agent pipelines and writes the Python code those pipelines run. The difficulty of these call sites varies by an order of magnitude, yet in practice a single model, chosen for the hardest site, serves all of …

BEACON-SP: Ontology-Grounded GraphRAG Framework for Clinical Suicide Risk Assessment

arXiv:2610.09026v1 Announce Type: new Abstract: We present BEACON-SP, an ontology-grounded Graph Retrieval-Augmented Generation (GraphRAG) framework for clinician-facing decision support in behavioral health settings such as suicide prevention, where effective assessment requires integrating heterogeneous clinical, behavioral, social, and temporal evidence. BEACON-SP combines patient knowledge graphs with ontology-gu…

Shared-Roadmap Generation and Evaluator for Multi-Agent Path Planning Using Heterogeneous Graph Neural Network

arXiv:2610.09034v1 Announce Type: new Abstract: Multi-agent path planning (MAPP) in continuous environments often relies on roadmaps to balance safety and search efficiency. However, traditional roadmap generation methods, such as lattice grids or standard sampling-based approaches, frequently face a trade-off between graph density and the likelihood of finding feasible, high-quality solutions. In this paper, we prop…

When the Governor Becomes the Disturbance: Control-Generated Disturbance and Cost-Aware Backoff in Governed Tool-Using Agents

arXiv:2610.09037v1 Announce Type: new Abstract: Supervisory governors can interfere with the tool-using agents they regulate. We study this possibility in a controlled file-recovery environment where increases in regulatory intensity trigger experimentally imposed tool failures. A cost-blind governor can turn these failures into persistent blocking that prevents task completion. We compare this governor with a backof…

From High Recall to High Utility: Dataset-Adaptive Post-Processing of LLM-Generated Customer Intents

arXiv:2610.09039v1 Announce Type: new Abstract: Large language models can extract useful signals from heterogeneous enterprise data, but high-recall extraction often produces outputs that are duplicated, uneven in granularity, semantically overlapping, or too numerous for downstream systems and human reviewers to use effectively. We present a dataset-adaptive post-processing architecture developed for Customer Intent…

Justice After Identity: Large Language Models and the View from Everywhere

arXiv:2610.09053v1 Announce Type: new Abstract: The search for a common view of justice and fairness has challenged human collective activity, as our diverging judgments are unavoidably shaped by the self-interests of social position, personal benefit, cultural inheritance, and historical circumstance. John Rawls famously attempted to overcome this limitation through popularizing a philosophical tradition known by th…

Epistemic Uncertainty-Aware Defect Detection for Quality Control in Medical Device Manufacturing

arXiv:2610.09057v1 Announce Type: new Abstract: Objective: We investigate whether accounting for epistemic uncertainty can improve the reliability of automated defect detection in medical device manufacturing. Methods: We consider a machine learning framework that operates on heterogeneous manufacturing and device-report data represented with Knowledge Graphs. To mitigate errors arising from uncertainty in the decisi…

RippleCP: Measuring Counterfactual Checkpoint Advantage in Coding Agents

arXiv:2610.09088v1 Announce Type: new Abstract: Agent checkpoint systems decide what state is recovery-relevant, how to snapshot it, and whether rollback is admissible. None decides which of the safe boundaries they expose are worth materializing. We formulate this as counterfactual checkpoint advantage, the reduction in future recovery cost obtained by checkpointing a candidate rather than skipping it, and measure i…

Constraint Tree Exploration for Learning from Language Feedback

arXiv:2610.09107v1 Announce Type: new Abstract: Natural-language feedback in interactive learning often explains why an action failed by pointing to violated requirements. Misinterpreting this feedback can lead an agent to rule out valid solutions. We study this setting by modeling user intent as latent constraints over an action space and formulating learning from language feedback as pure exploration over feasible …

GeoNatureAgent (GNA): A Framework and Benchmark for Pre-Production Evaluation of Tool-Using Agents on Geospatial and Environmental Tasks

arXiv:2610.09112v1 Announce Type: new Abstract: Before tool-using LLM agents are deployed in environmental and geospatial workflows, teams need evidence that an agent reliably selects the right operations against real APIs. We introduce GeoNatureAgent (GNA), a framework for pre-production evaluation of tool-using agents: a fixed sixteen-tool geospatial interface published as a Model Context Protocol (MCP) server, so …

From Uncertainty to Action: Learning to Steer LLM Agents

arXiv:2610.09115v1 Announce Type: new Abstract: Steering an LLM agent means deciding whether to correct it, at which step, and with which mechanism. Uncertainty is often used to decide when to correct an agent, but whether it can guide these decisions remains unclear. We steer agent trajectories separately at every non-terminal step with each of four mechanisms and run each continuation to completion. The resulting s…

Codex 与 ChatGPT Work 活跃用户达 4000 万新高，付费账户重置已全部到账

作者确认 banked reset 已到账所有账户，并转引 Day 3 动态称 Codex 与 ChatGPT Work 合计活跃用户达到 4000 万新高，其中提到 GPT-6 已在 Chat 中上线。

Build zero-trust AI agents with Google's Agent Development Kit

Building autonomous AI agents that mutate production state requires moving beyond soft system prompts to a robust zero-trust architecture. To secure Google Agent Development Kit (ADK) workflows against prompt injections and malicious execution, developers must implement hardware-backed cryptographic signatures for database writes, kernel-level sandboxing with gVisor for dynamic code, and deterministic semantic gatewa…

How to Evaluate Live & Voice Agents in ADK

Moving live voice agents from demo to production requires rigorous, automated testing to handle the unpredictability of real multi-turn conversations. ADK now provides native live evaluation, allowing developers to test graph-based agent workflows against LLM-driven simulated users that generate actual audio via Gemini TTS. By defining evaluation scenarios and natural-language rubrics, you can automatically score aud…

Decoding cosmic signals with deep learning and Keras

Astroparticle physics sits at the exciting intersection of astrophysics and particle physics and stu...

Enterprise-Grade Precision for Long-Context Multimodal Embedding Inference on Cloud TPU

Google Cloud has natively integrated TPU support into the vLLM serving engine, allowing developers to elastically scale high-demand embedding pipelines using Google Kubernetes Engine (GKE). To handle massive 15K+ token contexts for models like Qwen3-Embedding-8B, the engineering team implemented TPU-specific optimizations such as hardware-safe tensor alignment, JAX/XLA compilation pre-warming, and a hybrid StepPool a…

4 engineering patterns behind the strongest AI Agents Challenge submissions

The recent Google for Startups AI Agents Challenge revealed that the most successful multi-agent systems rely on foundational software engineering patterns rather than just raw model power. Winning architectures consistently implemented bidirectional MCP for seamless inter-agent communication, async event buses for parallel execution, strict unified validation for model fallbacks, and tiered routing to minimize expen…

Driving Developer Excellence: Inside the Program Sprints

The Gemini Enterprise Developer Experience (DevEx) program conducts ongoing sprint testing of end-to-end developer workflows to identify and rapidly resolve friction points without relying on internal shortcuts. This recent sprint focused on optimizing enterprise AI governance, including refining setup prerequisites, securing extension configurations, and clarifying policy enforcement mechanics to ensure a smoother, …

The Anatomy of Harness Engineering: How to Evaluate, Iterate, and Guard AI Coding Agents

While end-to-end benchmarks like SWE-bench provide broad performance scores for AI agents, they are often expensive, slow, and lack the root-cause diagnostics needed to explain exactly where an agent's logic broke down. To solve this, developers should adopt behavioral evaluations—fast, local, unit-style tests that assert on discrete intermediate actions, such as verifying specific tool calls or file modifications ra…

Announcing ADK for Kotlin 1.0: Building Production-Ready AI Agents in Kotlin, Android, and Beyond

Google has officially released version 1.0 of the Agent Development Kit (ADK) for Kotlin, achieving full feature parity with the Python and Java ADK cores to enable idiomatic, multi-agent AI development. Built on Kotlin Multiplatform (KMP), the framework leverages Kotlin Symbol Processing (KSP) for zero-reflection, type-safe function calling, alongside advanced orchestration capabilities like human-in-the-loop workfl…

Autonomous LLM post-training with Tunix on TPUs

The "autofinetune" project introduces an autonomous research loop that fully automates LLM post-training workflows, including Supervised Fine-Tuning (SFT) and Reinforcement Learning via GRPO. By defining boundary conditions and evaluation metrics in a single Markdown specification, developers can deploy an AI agent to iteratively edit training scripts, launch experiments, and automatically commit verified hyperparame…

Build zero-trust AI agents that judge intent, not just syntax

This blog post explores how to transition AI agents from static, build-time security controls to dynamic runtime governance using the Gemini Enterprise Agent Platform. It highlights three primary managed defenses: Model Armor for screening edge prompts, Semantic Governance Policies for evaluating tool intent against business rules, and Agent Anomaly Detection for catching multi-turn exploits. By shifting these capabi…

Agent Anomaly Detection, now in Private Preview on the Gemini Enterprise Agent Platform

Agent Anomaly Detection is a new, out-of-band oversight layer for the Gemini Enterprise Agent Platform that analyzes OpenTelemetry traces and tool calls to catch behavioral risks without adding runtime latency to live requests. It utilizes a multi-tiered detection pipeline—combining lightweight statistical scanning with deep LLM-based reasoning—to identify logical anomalies and policy violations grounded in the OWASP…

Why client SDK generation belongs in the open

Google has partnered with Speakeasy to open-source their OpenAPI code generation suite under the AGPLv3 license, a strategic move prompted by the sudden shutdown of Google's previous proprietary SDK provider. The newly open-sourced suite equips developers with deterministic, multi-language SDK generators that natively support strict typing and SSE streaming, alongside tools for compiling agent-native CLIs and documen…

Colab is now part of your Google AI plan

Unlock premium Google Colab compute with Google AI. Subscribers now get priority accelerators, Premium GPUs, and background execution for long training runs.

Introducing Support for Local AI Models in the Antigravity SDK

The Google Antigravity SDK now empowers developers to execute offline, agentic workflows locally using models like Gemma 4 26B A4B via LiteRT. This update facilitates powerful hybrid orchestration architectures, allowing a cloud model to act as a lightweight planner while local models securely handle token-intensive tasks—like code auditing and patching—directly on-device. Furthermore, the SDK provides drop-in suppor…

Turn your REST APIs into MCP tools with Google Cloud API Gateway

Google Cloud API Gateway now acts as a native remote Model Context Protocol (MCP) server, eliminating the need to build and maintain custom middleware to expose REST APIs to AI agents. By simply adding specific annotations (like x-google-api-management.mcp) to existing OpenAPI 3.x specifications, developers can instantly convert standard REST operations into discoverable, agent-ready tools. The gateway automatically …

Reproducing Olmo 3 7B Pre-training in MaxText: case study of large scale training on TPUs

The MaxText team successfully reproduced Ai2’s Olmo 3 7B language model from scratch on Google Cloud TPUs using JAX/XLA, precisely matching the original PyTorch-on-GPU reference across pre-training and mid-training stages on all held-out evaluations. The implementation achieved up to 57.4% Model Flops Utilization (MFU) and demonstrated robust infrastructure portability by surviving mid-run cluster resizes and cross-g…

Accelerating Spatio-Temporal Attention for Video Diffusion on TPUs

To address the quadratic latency bottleneck of self-attention in high-resolution video diffusion models, developers implemented Sparse VideoGen (SVG) to dynamically route attention heads to highly structured spatial or temporal sparse masks. Translating this algorithmic sparsity into physical hardware speedups on TPUs required optimizing the Splash Attention kernel by bypassing empty memory tiles, restricting exact c…

Bring multimodal semantic search to the edge with EmbeddingGemma 2

EmbeddingGemma 2 is a new 740M open-weight multimodal model that maps text, images, video, and audio into a unified vector space for privacy-first, on-device retrieval. Developers can easily integrate these capabilities cross-platform using MediaPipe Tasks or optimize fine-grained performance across CPU, GPU, and NPU accelerators with LiteRT. The model enables ultra-low-latency local solutions like search-as-you-type…

EmbeddingGemma 2: The Developer Guide

EmbeddingGemma 2 is a compact, open-source multimodal embedding model that maps text, code, images, video, and audio into a unified 768-dimensional space. Developers can use the sentence-transformers library to selectively load modular modality encoders—ranging from 270M to 740M parameters—to optimize memory usage. Additionally, Matryoshka Representation Learning enables dynamic dimension truncation down to 128d, sig…

Supercharge your development with the Google Developer Knowledge API ecosystem

Google is launching the Developer Knowledge API and MCP Server in public preview. This new toolset provides a canonical, machine-readable way for AI assistants and agentic platforms to search and retrieve up-to-date documentation across Firebase, Google Cloud, Android, and more. By using the official MCP server, developers can connect tools directly to Google’s documentation corpus, ensuring that AI-generated code an…

Quoting Ben Affleck

I've always been kind of into computers since I was young. And then when film started to move from analog film to digital, I became more interested in that aspect of it. And the visual effects workflow for many years has included machine learning. So I can write like pretty shitty Python scripts and stuff like that because with convolutional neural networks, which were the sort of precursors to what the transformer c…

Claude Haiku 5.5

As previously promised , here's Anthropic's new fast, low cost model: Introducing Claude Haiku 5.5 . The previous Haiku, 4.5, was very much showing its age. It came out almost a year ago , and was priced at $1/million input and $5/million output - relatively expensive even back then, and a full 10x the price of OpenAI's GPT-6 Luna , released last month. The new Haiku exactly matches the price of GPT-6 Luna - $0.10/$0…

GPT-6 Luna Decisions 上架 OpenRouter

OpenRouter 宣布 GPT-6 Luna Decisions 上线。OpenAI 的 Decisions API 可让应用选择合适的模型、工具或动作，支持发送文本、JSON 或图片并返回带概率的类型化答案。定价为输入 $0.10/M、输出免费，上下文 1M；引用 OpenAI 开发者账号称其决策速度比通过 Responses API 的 GPT-6 Luna 最快 10 倍。

Claude Haiku 5.5 in GitHub Copilot

Claude Haiku 5.5, Anthropic’s newest lightweight model, is now generally available in GitHub Copilot. It is designed for fast, high-volume work like subagents, quick edits, and terminal tasks. In early… The post Claude Haiku 5.5 in GitHub Copilot appeared first on The GitHub Blog .

LangChain 重构 Deep Agents 的 Skills 支持，新增工具绑定、固定技能与线程内重载

LangChain 重构 Deep Agents 的 Skills 支持，针对企业技能库增至数千个技能的场景推出三项更新：工具可绑定到技能、仅在该技能被读取时加载，用户可通过 /meeting-prep 之类的显式请求固定技能以在首次模型调用前加载，长线程可通过将 skills_metadata 设为 None 重载新增或变更的技能。

NVIDIA 与 Microsoft 推出 RTX Spark 平台并宣布 MXC 让 AI Agent 落地 Windows PC

NVIDIA 与 Microsoft 在旧金山活动上宣布为 Windows PC 共同打造 AI Agent 软硬件。

Cursor 公布 Claude Haiku 5.5 定价并下调 Claude Sonnet 5.5 缓存读取价格

Cursor 公布 Claude Haiku 5.5 定价为每 M 输入 token $0.10、输出 token $0.50，输入超过 100k token 时为 $0.50/M 和 $2.50/M。Claude Sonnet 5.5 缓存读取价格也从 $0.20/M 降至 $0.10/M，用户可在 cursor.com/evals 上通过 CursorBench 对比 Haiku 5.5 的表现。

Claude Code v2.1.293 发布：新增 Claude Haiku 5.5 并修复大量问题

Claude Code 发布 v2.1.293，新增 Claude Haiku 5.5（claude-haiku-5-5）作为 Anthropic API 默认 Haiku 模型，支持 1M 上下文，价格为 $0.10/$0.50 每百万 token（超 100K 提示为 $0.50/$2.50）。

Anthropic 为 Claude Max 和 Team 套餐推出月度 Platform API 额度

Anthropic 正在为 Claude Max 和 Team 套餐推出月度 Claude Platform API 额度：Max 5x 为 $100，Max 20x 为 $200，Team 最多 $500 且可共享。额度适用于任何模型，包括 Haiku 5.5，可在自己的代码或第三方 harness 中使用。

Anthropic 将 Claude Sonnet 5.5 缓存读取价格减半至每百万 token $0.10

Anthropic 宣布将 Claude Sonnet 5.5 的缓存读取价格减半，降至每百万 token $0.10。官方称这使 Sonnet 5.5 在多数长期运行任务上的运行成本降低约 20%。

OpenAI 向全部 ChatGPT 用户推出 GPT-6 与 Intelligent UI

OpenAI 发布面向更广泛用户的 GPT-6，并随 GPT-6 在 ChatGPT 中引入 Intelligent UI，可生成图形、按钮、表单、图表和可交互组件来回答问题。

Perplexity 开源 pplx-embed-v2-late 多模态 late-interaction 嵌入模型（9B 与 0.6B）

Perplexity 开源 pplx-embed-v2-late，两个针对文本和图像的 late-interaction 多向量嵌入模型，大小为 9B 和 0.6B，共享同一嵌入空间，权重已在 Hugging Face 提供。9B 可用于索引多模态数据，0.6B 可在设备端查询，无需 OCR 即可检索 PDF 页面；模型在 MADQA 得分 92.4%，BrowseComp+ 得分 64%。

Unsloth 开源教程：本地训练 Qwen3.5 0.8B 决策模型，准确率从 20.7% 提升至 74.3%

Unsloth 发布教程与开源仓库，可将 Qwen3.8、Gemma 4 等 LLM 微调为输出选项概率的决策模型，Qwen3.5 0.8B 在 3 个决策基准上的合计准确率从 20.7% 提到 74.3%，仅需 4GB 显存。

Purpose-built model for leaked secret detection

Secret protection should keep pace with the way you build software, whether you write code yourself or work with an AI agent. With our new purpose-built model, we’re bringing context-aware… The post Purpose-built model for leaked secret detection appeared first on The GitHub Blog .

Microsoft Research Asia 开源 Agent Lightning v1.0：3,500 行代码的真实 harness 智能体 RL 训练框架

Microsoft Research Asia 提出 Harnessed Agentic RL 训练范式并开源重建的 Agent Lightning v1.0，让部署时使用的同一 agent harness 直接参与强化学习，无需在训练框架内重写 agent。

Local sandboxing for GitHub Copilot now generally available

Local sandboxing for GitHub Copilot is now generally available in GitHub Copilot CLI, the GitHub Copilot app, and VS Code sessions using Agent Host. Local sandboxes give developers a secure… The post Local sandboxing for GitHub Copilot now generally available appeared first on The GitHub Blog .

Discover local models in GitHub Copilot CLI

GitHub Copilot CLI makes it easier to choose a local model without leaving your existing workflow. Starting in CLI version 1.0.94-0, use /model to discover supported models from a running… The post Discover local models in GitHub Copilot CLI appeared first on The GitHub Blog .

亚利桑那州法院裁定 AI 生成受害者视频带有不当情感分量 将重新量刑

亚利桑那州上诉法院裁定，Gabriel Horcasitas 2021 年路怒枪杀 Christopher Pelkey 一案维持过失杀人定罪，但因量刑听证中播放的 AI 生成受害者视频带有不当情感分量，法官须重新考虑刑期。

Google 开放 SynthID Detector 门户，可检测图片、视频和音频是否由 AI 生成

Google 介绍检测媒体是否由 AI 生成的方法：访问 https://synthid.com 上传图片、视频或音频文件，门户会扫描文件是否包含来自 Google 或其合作伙伴的 SynthID 水印。

Can a Cloud-Native Harness Make Agents Reliable Beyond the Desktop?

Kubernetes co-creators Craig McLuckie and Joe Beda aim to bring agent harnesses fully into the cloud.

a16z 解析德州为何让数据中心排队等电

a16z 的 Ryan McEntush 分析德州电网暂停审批数据中心的原因：并网队列从 2024 年底的 63 GW 激增到今年 6 月的 474 GW，约 90% 是数据中心，开发商大量投机性申请且社区沟通不足。

Nemotron 系列微调后达到 IOI 2026 与 IMO 2026 金牌水平

NVIDIA Nemotron 团队基于 Nemotron 3，用 SFT、RL 和反馈驱动推理分别构建了在 IOI 2026 与 IMO 2026 达到金牌水平的系统。

Helping teens learn, plan, and shape the future of AI

College Planner is coming to ChatGPT for Teens to help students manage college applications, alongside new flashcards, quizzes, and a teen AI council.

Radisson Hotel Group brings hotel discovery into ChatGPT

Radisson partnered with Accenture to build a ChatGPT plugin using OpenAI technology, helping travelers find, compare, and book hotels while planning their trips.

Mistral Large 4 进入 Code Arena: WebDev 排名第45，得分1534

Arena 宣布 Mistral Large 4 登陆 Code Arena: WebDev，以 1534 分排名第 45，比 Mistral Large 3（第130名）高 304 分，比 Mistral Medium 3.5 高 271 分。

[AINews] Quasi-Riemann-Hypothesis: OpenAI publishes 722 math papers solving 90 of the top 500 open math problems; “the most significant moment” in >100 years of mathematics

Our head hurts.

Quoting Jake Boggan

I was a graph theory junkie long ago and even moved to Budapest for awhile to study among the greats. While I was there I started working on Barnette's Conjecture which came to occupy my thoughts over the next 24 years of my life, on and off as I worked in many different fields. Last summer I even thought for a few days that I had actually solved it. But it's supposedly proven here - problem 180 . I don't know what t…

OpenAI “rogue” agent activities found on Wikimedia projects

OpenAI “rogue” agent activities found on Wikimedia projects Given how tempting a target wikis are for rogue agent swarms, it's not a huge surprise that Wikipedia found evidence of that activity once they went looking: The Wikimedia Foundation conducted its own investigation to see whether Wikimedia websites had been similarly affected by AI agents, focusing on those operated by OpenAI. We can confirm that we have dis…

HeyGen x Google Cloud: Bringing Avatar IV to TPUs

HeyGen ported their 18B+ parameter Avatar IV video generation model to Google Cloud's Trillium (v6e) TPUs via torchax and XLA, utilizing FSDP and Ulysses sequence parallelism across an eight-chip mesh. To achieve a 1.86x speedup for real-time streaming, the engineering team pipelined exposed all-to-all collectives, aligned sparse attention block sizes to eliminate mask padding, and bypassed softmax serial dependencie…

Google Research 三个月专利起草实验发现AI辅助未必能培养初级律师的专业判断

Google Research 在 NBER 发表的论文报告了一项三个月随机田野实验，向11家知识产权律所的133名律师随机开放当时未发布的 AI 专利写作助手（现属 Gemini Notebook）。

Google 发布 Developer Knowledge API 生态，为 AI 智能体提供官方文档检索

Google 推出 Developer Knowledge API 生态，作为 Google Cloud、Firebase、Android 等开发者文档的官方程序化来源，用结构化 API 和 Markdown 格式文档取代网页抓取，支持语义与关键词搜索、文档分块和有依据的问答。

Liquid AI 发布开源决策模型 d1-3B 和 d1-omni-600M

Liquid AI 发布开源权重决策模型 d1-3B 和 d1-omni-600M，已在 Hugging Face 上提供。

vLLM 详解 DeepSeek-V4.1-Flash 优化：Agent 场景吞吐提升 5 倍

Inferact 与 vLLM 社区在 DeepSeek-V4.1-Flash 发布三周内完成优化，低并发速度提升 1.9 倍，150 TPS 约束下吞吐提升 5.3 倍。

GPT-6 and Intelligent UI for everyone

GPT‑6 is rolling out globally in ChatGPT with Intelligent UI, delivering faster responses with visuals and interactive experiences you can explore and use directly.

Quoting Victoria Kim

Since the Medicare breach, OpenAI has put in place additional monitoring to allow “immediate intervention” by staff to stop training if the company’s models access the internet in ways they’re not supposed to, Mr. Kwon [chief strategy officer at OpenAI] said. — Victoria Kim , Reporting from the Australian parliament Tags: accidental-cyberattacks , generative-ai , ai-security-research , openai , ai , llms

Update your IDE to restore agent activity in Copilot usage metrics

If your Copilot usage metrics have shown agent activity or agent lines of code falling while Copilot usage kept growing, we’ve found the cause, and a fix is rolling out… The post Update your IDE to restore agent activity in Copilot usage metrics appeared first on The GitHub Blog .

llm-openai-decisions 0.1a0

Release: llm-openai-decisions 0.1a0 OpenAI released their new Jev-style Decisions API , as previously announced at last week's DevDay. Since I already have an llm-typesafe plugin for talking to Jev, I had GPT-6 Astra read the new OpenAI API documentation and build an llm-openai-decisions plugin inspired by llm-typesafe . Unlike Jev, the new gpt-6-luna decision model supports image input in addition to text. Both mode…

OpenAI 发布内部前沿模型产出的数学研究成果

OpenAI 发布一批由内部前沿模型产出的新数学成果，成果以 GitHub 仓库形式公开，并附论文修订与引用协议，其中许多证明已用 Lean 形式化以便计算机验证。

llm-mistral 0.16

Release: llm-mistral 0.16 Adds support for reasoning models, such as the newly released Mistral Large 4 . Tags: llm , mistral , llm-reasoning

GitHub 重建 Git 基础设施，应对智能体规模开发

GitHub 宣布重建 Git 基础设施，以支持智能体规模开发带来的高并发读写负载。2026 年 8 月 GitHub 月度 Git 事件量达 473.3 billion（一年翻倍以上），9 月智能体和开发者产生 7.38 billion commits（超一年前五倍），pushes 同比增长 4.9 倍。

ChatGPT 推出 Meetings 插件，可自动记会议纪要并跟进待办

ChatGPT 发布 Meetings 插件，可替用户记会议笔记，并基于 ChatGPT 对用户和既往工作的了解，在 ChatGPT Space 保存个性化摘要和后续步骤。笔记可私密保存或与团队共享，还能让 ChatGPT 更新项目计划或起草跟进内容。目前以 beta 形式面向 Pro 和 Business 用户开放，可在 macOS 桌面应用的插件目录搜索 Meetings 使用，Enterprise 版即将推出。作者 Tibo 补充称，可放心畅谈并将积累的上下文载入 codex 来完成任务。

EmbeddingGemma 2

My comment on EmbeddingGemma 2 — Hacker News. I really appreciate that EmbeddingGemma 2 is under the Apache 2.0 license. For embedding models in particular, I don't think it makes sense to use a closed, proprietary, hosted-only model. Most applications of embedding models involve calculating thousands or even millions of embedding vectors and storing them for later comparison. If your model is proprietary, the …

Introducing Mistral Large 4: Le chonk

Introducing Mistral Large 4: Le chonk Mistral are back in the game. Today they're releasing a preview of Mistral Large 4, a 1 trillion parameter, 49 billion active parameter model trained on their own cluster of 3,800 NVIDIA Grace Blackwell GPUs. The preview is available via their API. They promise to release the open weights model at the "end of this month". The model only supports two reasoning levels - "none" and …

Stacked pull requests generally available

GitHub stacked pull requests are now generally available. Break large changes into smaller, focused pull requests that you can review independently and merge together. Since the feature went into public… The post Stacked pull requests generally available appeared first on The GitHub Blog .

EmbeddingGemma 2: an open, lightweight multimodal embedding model

Google DeepMind 发布开源轻量多模态嵌入模型 EmbeddingGemma 2

Google DeepMind 发布 EmbeddingGemma 2，基于 Gemma 4 架构，以 Apache 2.0 许可开源，将文本、代码、图像、视频和音频映射到统一嵌入空间。

亚利桑那州法院裁定AI生成受害者视频带有不当情感分量，凶手须重新量刑

亚利桑那州上诉法院裁定，Gabriel Horcasitas过失杀人案量刑中使用的受害者Christopher Pelkey AI生成视频带有不当情感重量，罪名维持但刑期须重新考虑。

Using Parseable with Datasette for OpenTelemetry traces

TIL: Using Parseable with Datasette for OpenTelemetry traces I saw Parseable in a Show HN today - it's a new observability platform with both an open source (AGPL) Rust implementation (a single ~180MB binary), an "Enterprise" version with extra features and a cloud hosted option. Since Datasette 1.0a41 added OpenTelemetry support (thanks, Alex Garcia), I decided to fire up Codex and have it figure out how to run Pars…

Mistral Large 4

My comment on Mistral Large 4 — Hacker News. wren6991 : The benchmark is saturated. Frontier models are tested with an armadillo in fishnet tights jaywalking on Mars. OK well I couldn't resist this one: llm -m claude-opus-5.5 'Generate an SVG of an armadillo in fishnet tights jaywalking on Mars' llm -m gpt-6.1-sol 'Generate an SVG of an armadillo in fishnet tights jaywalking on Mars' llm -m gemini-3.8-flash 'Ge…

datasette-atom 0.11a0

Release: datasette-atom 0.11a0 A minor fix for compatibility with the latest Datasette alphas. This meant we could upgrade the datasette.io site to Datasette 1.0a41. Tags: atom , datasette

DeepSeek 据报道接近完成至少 800 亿元融资，腾讯与宁德时代参与

据 Bloomberg 援引知情人士报道，DeepSeek 接近完成至少 800 亿元（约 120 亿美元）融资，高于原定约 500 亿元的目标。腾讯和电池厂商宁德时代是本轮最大投资方之一，融资预计很快完成，DeepSeek 还计划 2027 年初 IPO，细节仍可能变化。

Sierra 与 Meta 联合多家企业发布 Personal Agent Protocol 开放协议

Sierra 与 Meta 联合 Genesys、Instinct、Rocket、Shopify、Stripe、Walmart 等伙伴宣布开发 Personal Agent Protocol，一个定义个人 AI 智能体如何与企业交互的开放标准，任何人都可实现。

Atlassian and OpenAI expand partnership to turn enterprise knowledge into action

Atlassian and OpenAI are expanding their partnership to connect frontier models with enterprise knowledge and help teams plan, build, and deliver work.

Anthropic 发布 Claude Haiku 5.5，Artificial Analysis 评测得分 43

Anthropic 发布 Claude Haiku 5.5，在 Artificial Analysis Intelligence Index 得 43 分，较上一代 Haiku 一年内提升 26 分，是首个支持 effort 设置与 adaptive thinking 的 Haiku 模型。

Scrimshaw Jukebox

Tool: Scrimshaw Jukebox I wanted to see if Claude Opus 5.5 could compose music, so I tried this : I want you to write some computer game music for me. First design simple text based format for the music and build an artifact that can play it out loud - include some example tracks in that artifact I am looking for music of the quality of the original secret of Monkey Island It leaned a lot harder into the Monkey Islan…

The Cyber Risk Discourse is Broken

Open-weights, ideology, and acknowledging trade-offs.

AIGC Weekly #191

欢迎收看第 191 期 AIGC 周刊

How Jump Trading is scaling quant research with ChatGPT

Jump Trading uses OpenAI to expand quantitative research. See how longer-running AI workflows combine multiple data sources with human review.

Sharing AI progress in mathematics

OpenAI publishes new results on open problems in mathematics from an internal frontier model and shares Lean proof formalizations and research details on GitHub.

Code scanning AI Scan enablement status in security overview

Organization and enterprise administrators can now see AI Scan for pull requests enablement status in the security overview coverage view. The code scanning summary shows enabled and not enabled repository… The post Code scanning AI Scan enablement status in security overview appeared first on The GitHub Blog .

Advancing computer use with Ironclad

Learn how OpenAI and Ironclad are training and evaluating AI agents on complex contracting workflows to advance computer use for professional work.

[AINews] Reflection Beam - 501B-A23B American Open Model

A small win for US open source

Anthropic 5180亿美元计算力支出中约4137亿美元无论使用与否都需支付

Anthropic 计划未来数年在云计算和计算力上支出 5180 亿美元，其中约 4137 亿美元为不可撤销承诺，即使容量闲置也需支付，平均每年约 410 亿美元。

Introducing Credentio: Open Source C++ Library for C2PA Content Credentials from Google

Credentio is a newly released, open-source C++ library from Google that allows developers to integrate high-performance, local-first validation of C2PA Content Credentials into their client and server applications. By processing assets entirely locally with a highly optimized memory footprint, the library delivers instant validation verdicts for multi-gigabyte media files without incurring cloud latency, bandwidth co…

Why Go is an Ideal Language for AI-Assisted Software Engineering

As AI coding assistants shift the developer's primary role from writing boilerplate to reviewing and maintaining systems, language choice becomes critical for long-term architectural integrity. Go directly addresses this new paradigm by utilizing its strict compiler, integrated toolchain, and uncompromising readability to provide deterministic guardrails that help AI models self-correct and generate highly standardiz…

Quoting Felix Rieseberg

The "old" version of Cowork runs model inference in the cloud, executing tool calls in an Anthropic-provided VM we shipped to your computer. We added the VM for capability, safety, and security reasons - mapping in just the data you explicitly added to your session. People loved what they were able to do with Claude but didn't love the disk, battery, and performance cost of running the VM locally. Also, people didn't…

Secret scanning adds detectors for Lovable, Supabase, and more

Secret scanning now detects new secret types from Lovable Labs, Pydantic Services Inc., and Supabase. New secret scanning partner The following provider joined the secret scanning partnership program. When one… The post Secret scanning adds detectors for Lovable, Supabase, and more appeared first on The GitHub Blog .

Our approach to EU text provenance rules

How OpenAI is approaching text watermarking under EU rules. Learn where watermarks apply, how detection works, and why access starts with researchers.

Building advertising for the way people use AI

OpenAI introduces a new visual ad format in ChatGPT and expands measurement tools, attribution partnerships, and brand suitability for advertisers.