Summary of AI Updates

Real-Time Intelligence with IBM Time Series Models on Confluent

Confluentに統合されたIBM Time Series Modelsによるリアルタイムインテリジェンス

MrBeast partners with Gemini to turn impossibly big ideas into reality

MrBeastがGeminiと提携し、大胆なアイデアを現実に

Mr. Beast

ATV Big Air Tour turned 3 days of work into 3 hours with ChatGPT

ATV Big Air Tourが3日間の作業をChatGPTで3時間に短縮

ATV Big Air Tour uses ChatGPT Work to speed up marketing, merchandising, and more. It even turned merchandise photos into an inventory website in 15 minutes.

Bilevel Coordinated Reflection: A Game-Theoretic Approach to Multi-Agent LLM Systems

二段階協調反射：マルチエージェントLLMシステムへのゲーム理論的アプローチ

Multi-agent LLM systems commonly use an orchestrator to decompose a task for a team of workers and then improve through textual reflection. Despite strong empirical results, these systems lack a unified account of coordination, memory improvement, and the role of external verification. We model orch…

RoboTok: An Internet-Scale Data Engine for Human Demonstration Retrieval and Dexterous Manipulation Learning

RoboTok: 人間のデモンストレーション検索と器用な操作学習のためのインターネット規模データエンジン

Robot learning increasingly depends on broad and diverse demonstrations, yet collecting robot data remains expensive and poorly suited to covering the long tail of real-world tasks. To address this bottleneck, we introduce RoboTok, an internet-scale data engine that, given a query human manipulation…

Language Models Can Control Their Own Attention

言語モデルが自身の注意機構を制御可能に

Language models spend most of their attention on a small fraction of context, yet they read the entire KV cache to find the few tokens that matter. If the user asks about a previous detail in a 1M-token conversation, global attention layers must scan the full context to generate each token of the re…

EarlyEval: Cheaper Agent Evaluation via Early Outcome Prediction

EarlyEval: 早期結果予測によるエージェント評価の低コスト化

Evaluating LLM agents is essential for guiding their development, yet it has grown prohibitively expensive: a single pass of a frontier model over an agentic benchmark can cost hundreds to thousands of dollars, a price paid repeatedly across iterative development cycles. Prior efforts, centered on b…

SolarWM: Open Data and Scalable Training for Long-Horizon Video World Models

SolarWM: 長期予測ビデオワールドモデルのためのオープンデータとスケーラブルな学習

We introduce SolarWM, a fully open foundation for building interactive video world models from data preparation through long-horizon inference. Training across heterogeneous data sources and video backbones is challenging: datasets differ in temporal scale, camera geometry, visual quality, motion, a…

Repo-To-Skill: Distilling GitHub Repositories Into AI4AI Skills

Repo-To-Skill: GitHubリポジトリをAI4AIスキルに蒸留

Autonomous agents are beginning to carry out machine-learning (ML) research end to end. These agents combine a model backbone with a harness for planning, execution, memory, and verification, but this architecture still leaves domain-specific know-how outside the agent. We call this missing layer op…

Gemini API リリースノート 2026-09-02

Gemini 3.8 Flash generally available (GA) : Released gemini-3.8-flash , our most intelligent Flash model, engineered for long-horizon software engineering, autonomous agents, and complex enterprise workflows. To get started, see the Gemini 3.8 Flash model page and the Latest model guide .

OpenAI API Changelog 2026-09-02

OpenAI APIチェンジログ 2026-09-02

Update Updated API errors so applications can distinguish traffic that increases too quickly from temporary model overload. Traffic that increases too quickly can return a 429 error with the slow_down code. Temporary model overload returns a 503 error with the server_is_overloaded code. Both respons…

BenchMIRT: What are LLM benchmarks actually measuring?

BenchMIRT: LLMベンチマークは何を実際に測定しているのか

The latest AI news we announced in August 2026

2026年8月に発表したAI関連ニュース

Transitioning cards: 1. Text "Gemini 3.7 Flash" next to the Gemini logo icon; 2. a photo of a pixel phone; 3. Google Gemini logo above the text "Claim your student plan for 1 year at no cost"

Introducing agentic video understanding with Gemini

Geminiのエージェンティック動画理解機能を発表

Text "Agentic video understanding" next to the Gemini logo, all on a dark blue background

How AI-native companies turn workflows into operating capability

AI企業がワークフローを運用能力に変える方法

Basis, Clay, and Exa Labs use AI agents to improve onboarding, account management, and developer integrations. See what enterprise leaders can apply.

Try Google Pics: Easy image creation and editing in Google Workspace

Google Picsを試す: Google Workspaceで簡単な画像作成・編集

Collage of images created by Google Pics, with the text "Say hello to Google Pics" on top

Path to Astra: critical capabilities and frontier safeguards

Astraへの道: 重要な機能とフロンティアセーフガード

Astra is the first OpenAI model to meet the Critical cybersecurity capability threshold under the Preparedness Framework, with stronger safeguards for release.

Healthcare organizations can now connect EHR and additional industry data to ChatGPT

ヘルスケア組織がEHRと業界データをChatGPTに接続可能に

ChatGPT can now connect to trusted healthcare data, helping clinicians securely access patient context, medical research, and more.

How law firm Gilbert + Tobin governs and scales AI with OpenAI

法律事務所Gilbert + TobinがOpenAIでAIを統治・スケーリング

See how Gilbert + Tobin combines CEO-led commitment, rigorous governance, and human accountability to scale ChatGPT Enterprise and Codex across the firm.

LatentPress: Context Compression Beyond Text and Vision

LatentPress: テキストと画像を超えたコンテキスト圧縮

Compressed context is usually carried as human-readable text or as rendered images that must be decoded, even when its consumer is a language model. We introduce LatentPress, which writes conversational histories and long documents into a third representation: continuous memory tokens that a frozen…

HarnessDev: Can LLMs Create and Evolve Their Own Agent Harness?

HarnessDev：LLMは独自のエージェント Harnessを構築・進化させられるか

As agents move from research prototypes to deployed tools, their capability increasingly depends on model-external execution infrastructure, commonly termed the agent harness. Changing this harness while holding model weights fixed can substantially alter task performance. Current agent evaluations…

OpenAI API Changelog 2026-09-01

OpenAI API Changelog 2026-09-01

Update Connections to api.openai.com can now use IPv6.

It Takes Two to Match: Co-Evolving Generative Retriever with Reinforcement Learning

It Takes Two to Match: 強化学習による生成型レトリーバーの共進化

Retrieval is the first stage of modern search and advertising systems, selecting a candidate set from a large item universe for downstream ranking and auction. Recent work increasingly leverages LLMs to improve retrieval through query expansion, data synthesis, and retrieval-feedback training. Howev…

From Production Traffic to Post-Training: Building a Self-Hosted LLM That Covers the Corporate Request Mix

本番トラフィックからポストトレーニングへ：企業のリクエスト混合に対応するセルフホストLLMの構築

Data-residency constraints force enterprises to self-host LLMs, but continuous adoption of newer models without decommissioning their predecessors expands the serving fleet, fragmenting a finite GPU pool. We consolidate traffic from over 200 internal applications onto a single model by closing quali…

H3-World: Turning Language Understanding into World Control

H3-World: 言語理解を世界制御へ

We present H3-World, an efficient framework that turns the 33B MiniMax-H3 video generator into an interactive world model. Our key finding is that, as large video generators become more capable, language is emerging as a natural interface for control. MiniMax-H3, for example, already supports zero-s…

SMELT: Scaling Laws for Compute-Matched MoE Looped Transformers

SMELT: コンピュート規模を揃えたMoEループトランスフォーマーのスケーリング則

Looped Transformers increase effective depth by iterating a shared block of layers, but most evaluations compare at fixed model size, conflating architectural advantage with extra FLOPs. We study looping on Mixture-of-Experts Transformers while closely matching per-token FLOPs, total non-embedding p…

StudentSim: Training LLM-based Student Simulators

StudentSim: LLMベースの学生シミュレータの訓練

AI tutors are most useful when they adapt to each student's strengths, weaknesses, and preferred guidance, but evidence about which guidance works for which student is sparse, slow, and costly to collect from real learners. Student simulators can provide this signal as a proxy, yet existing approach…

Introducing @huggingface/kernels: 200+ WebGPU Kernels for Local AI

@huggingface/kernelsの発表: ローカルAI向けWebGPUカーネル200+

Gemini API リリースノート 2026-09-01

Agentic video understanding : Released agentic video understanding for Gemini 3.7 Flash, 3.6 Flash, and 3.5 Flash-Lite across the Interactions and GenerateContent APIs. The model dynamically navigates video timelines, requesting transcripts, frames, or audio tracks on demand. This approach uses up t…

Developing Enterprise Frontier Safeguards with our customers

顧客とのエンタープライズフロンティアセーフガード開発

PFNのマウス尾静脈自動注射ロボットAUTivが日本ロボット学会「実用化技術賞」を受賞

PFN、PCゲーム『Omega Crafter（オメガクラフター）』プレイヤー数10万人突破を記念し40%割引セールを実施

PFN 2028年入社 新卒採用 プレエントリー受付開始

Introducing Azure Multicloud Interconnect for AWS

AWS向けAzure Multicloud Interconnectを発表

Azure Multicloud Interconnect helps simplify private connectivity between Microsoft Azure and AWS, enabling organizations to support multicloud and AI workloads with a more streamlined, cloud-native networking experience. Built on open interoperability standards, it helps reduce operational complexi…

Inside Microsoft’s marketing team: Scaling expertise with AI

マイクロソフトのマーケティングチーム内幕：AIで専門知識をスケーリング

AI is helping organizations meet high expectations as markets change quickly and technology advances at a rapid pace. The post Inside Microsoft’s marketing team: Scaling expertise with AI appeared first on Microsoft Azure Blog .

Polimill builds Japan's next-generation public AI infrastructure

Polimillが日本の次世代公的AI基盤を構築

Polimill uses OpenAI GPT models and Codex to help municipalities search and use administrative knowledge while accelerating development.

OpenAI supports California’s bill to advance youth AI safety

OpenAIがカリフォルニア州の若年層向けAI安全法案を支持

OpenAI supports California SB 1119, advancing strong, age-appropriate AI safeguards for teens while preserving opportunities to learn, create, and explore.

A milestone in expanding access to AI

AIへのアクセス拡大における重要な節目

ChatGPT Ads reaches $1 billion in annualized revenue run rate and expands globally, supporting broader access to AI through free and affordable options.

Dr. Claw: An AI Scientist Workspace for Vibe Research

Dr. Claw: バイブリサーチ向けAIサイエンティストワークスペース

Command-line coding agents (e.g., Claude Code, Gemini CLI) can already read and write files and sustain long sessions, yet end-to-end research still fragments across chat tools, IDEs, terminals, and writing environments, and the decisions that make it auditable are rarely preserved. We present Dr. C…

Aspire: Can Models Self-Evolve from Vague Goals?

Aspire：モデルは曖昧な目標から自己進化できるか

Many important forms of human learning begin with a vague goal, such as "become a better physicist" or "improve at research." Learners must interpret the goal, identify capability gaps, decide how to learn, and determine whether they have actually improved. In contrast, existing work on LLM self-evo…

ZimaBlue: Evolving Generalizable World Action Models through Scalable Video Pre-training

ZimaBlue: スケーラブルなビデオ事前学習を通じた汎用世界行動モデルの進化

Robotic manipulation faces a fundamental scaling challenge: robust generalization demands broad physical experience, yet action-labeled robot trajectories are expensive to collect and inherently limited in diversity. Egocentric videos offer a far more scalable source of embodied experience, capturin…

Qwen-Drive-1.0: An Initial Step towards a Vision-Language Foundation Model for Autonomous Driving

Qwen-Drive-1.0: 自動運転向けビジョン言語基盤モデルへの第一歩

We present Qwen-Drive-1.0, an initial step towards a vision-language foundation model for autonomous driving. Qwen-Drive-1.0 retains the architecture of the pretrained vision-language model (VLM) and integrates 3D perception, visual question answering, and motion planning within a unified framework.…

On the Design of Qwen3.8-Next Architecture: Evaluation, Efficiency, and Training Stability

Qwen3.8-Nextアーキテクチャ設計について: 評価、効率性、訓練の安定性

We describe the architecture and ablations of Qwen3.8-Flash-Next, a sparse mixture-of-experts model with 125B parameters, 6B activated per token, and additional 51B parameters of n-gram embedding tables held off the accelerator. On fourteen pre-training benchmarks the model leads the 397B-A17B prede…

PaperGym: Rubric-Centered Evolution for Research-Plan Generation

PaperGym: 研究計画生成のためのルーブリック中心進化

Research planning is the decisive capability of AI scientists. Yet a research plan admits no verifiable answer, so reinforcement learning lacks the environment it requires: tasks paired with a critic. Rubrics extracted from scientific papers can supply the critic. Existing pipelines, however, draw t…

Normalized Low-Rank Adaptation

正規化低ランク適応

While low-rank adaptation (LoRA) is widely used for parameter-efficient model adaptation, how to regularize its training dynamics for stable and effective optimization remains underexplored. Because LoRA initializes the up-projection to zero, its early optimization dynamics are largely governed by t…

Lucida: Parse, Generate, and Place for Composable Real-to-Sim Scene Modeling

Lucida: 合成可能なReal-to-Simシーンモデリングの解析・生成・配置

Composable scene modeling aims to recover a real indoor scene as complete, editable object assets arranged as observed, giving robot simulation and embodied AI a simulation-ready replica of the real environment whose objects can be manipulated individually. Existing pipelines decompose the task into…

Does On-Policy Distillation Really Distill? From Noisy Teacher to Self-Improvement

オンポリシー蒸留は本当に蒸留しているのか: ノイジーな教師から自己改善へ

On-policy distillation (OPD) offers dense token-level supervision as an alternative to the sparse outcome-level advantages of reinforcement learning with verifiable rewards (RLVR). However, the teacher scores student-generated trajectories that are inherently off-policy for it, so the reliability of…

DreamX-Creator: Democratizing Native Audio-Video Generation at 2K Resolution

DreamX-Creator: 2K解像度のネイティブ音声動画生成を民主化

Recent video generators often omit audio or synthesize it in a separate stage, limiting reciprocal modeling of visual dynamics and acoustic events. We present DreamX-Creator 1.0, a compact native joint audio-video generation system centered on a 7B generator. Conditioned on a first frame and a text…

Improving our alignment and security efforts

アライメントとセキュリティ対策の強化

Scaling Automatic Research Agents via World Models

ワールドモデルを活用した自動研究エージェントのスケーリング

Automating empirical research is a long-standing direction of AI. Recent automatic research (AutoResearch) agents bring this goal within reach, as modern LLMs show the capability to independently implement solutions and learn from the execution outcomes. Behind these gains, post-training (especially…