AI各社の公式発表を1画面に。各社のページを個別に開かなくても最新の動きが追えます。
ExplorationBench: 検証可能な異世界でのAIシステムの探索能力を測定
Scientific discovery begins where known problems end. There, AI systems must engage in exploration: framing hypotheses, designing experiments, and iterating on the results. However, evaluating this ability is difficult: (1) how to verify whether a genuinely new hypothesis holds, and (2) how to deter…
Rufus-Air: オープンLLMのファインチューニングレシピ
Rufus-Air is an open and reproducible post-training recipe on GLM-4.5-Air-Base (106B-A12B), organized as a serial pipeline of eight stages: SFT, Reasoning RL, Coding RL, Instruction-Following RL, General Agent, Coding Agent, Search Agent, and RLHF. We document the data, reward design, infrastructure…
WanPE：現代のテキスト・トゥ・ビデオ生成向け映画的プロンプト強化
Video generation begins in text space by authoring a cinematic screenplay, then materializes into pixels. As contemporary video generators scale to 30 seconds and faithfully follow complex conditions, the textual prompt largely directs the production, planning how actions, camera trajectories, light…
Transformerは2つの思考を同時に保持可能：LLMの線形重ね合わせの証拠
While Large Language Models (LLMs) rely on highly non-linear components, in this work we demonstrate that they exhibit fundamental linearity: when inputs from distinct text streams are linearly combined, the model outputs a superposition of the individual next-token distributions. We term this the S…
次なるフロンティア：AI先駆者Jürgen Schmidhuberを Sakana AIが迎える
NVIDIA Warp と MjWarp でロボティクスシミュレーションと学習ワークフローを高速化する方法
Google Beam、新地域、新パートナー、新顧客で拡大
Google Beam promotional animation
安全なサーバーサイドメモリでプライベート AI コンピュートを推進
Introducing private, server-side memory to Private AI Compute for personal AI.
新世代の Connected Apps が Gemini に展開される
Gemini is adding new Connected Apps, including Adobe, Airtable, Linear, Peloton and more to help you easily tackle your to-do list.
OpenAI Academy、2年の歩み
Marking two years of OpenAI Academy and bringing AI skills to even more communities.
Gemini 3.8 テキスト音声変換、登場
a text card image reading "Introducing Gemini 3.8 Flash TTS and 3.8 Flash-Lite TTS"
エージェント中心のプラットフォーム設計：エージェントが作業を担当する場合の変化
The organizations pulling ahead are not simply adding AI to what they already have. They are designing for a different kind of software. The post Designing agent-first platforms: What changes when agents do the work appeared first on Microsoft Azure Blog .
OpenAIがウクライナに民間防衛向けのサイバーアクセスを提供
OpenAI is extending access to its Daybreak program to the Government of Ukraine to support the cyber defense of civilian infrastructure.
RinggのAIエージェント、OpenAIで顧客対応の65%を解決
Using GPT-5.6, Ringg powers multilingual agents across voice, chat, WhatsApp, and web for 90% less cost vs. GPT-4.1.
invideoがGPT-6 Astraで色補正を3倍向上
With GPT‑6 Astra, invideo plans edits with greater precision, improves color correction and grading threefold, and produces 50 custom effects in one day.
HarveyがGPT-6 Astraで法的文脈を活かし、より強い文案を生成
GPT-6 Astra produces more structured, context-aware legal documents, freeing lawyers to focus on strategy.
サム・アルトマンの国連安全保障理事会での発言
OpenAI CEO Sam Altman discusses AI safety, human control, and international cooperation in remarks to the United Nations Security Council.
MentalHealthBenchを紹介
MentalHealthBench is an expert-informed benchmark for evaluating helpful and safe AI responses across realistic mental health conversations.
ChatGPT広告が東南アジアと台湾に拡大
ChatGPT Ads is expanding to Southeast Asia and Taiwan, giving eligible businesses new ways to reach people across more than 60 countries.
AirbnbがGPT-6 AstraおよびOpenAI frontier modelsへのアクセス拡大
Learn how Airbnb is expanding access to GPT-6 Astra and OpenAI frontier models to help engineering teams solve bugs, design systems, and ship faster.
Agent-Editing World Model: LLMエージェント向け世界モデリングの再考
Recent advances in large language models (LLMs) have enabled agents to tackle long-horizon tasks across diverse environments. To further improve agent performance, existing language world models typically predict environment observations, yet reconstructing high-entropy, execution-dependent tool res…
ワールドモデルにおけるオブジェクト永続性の学習
Object permanence and solidity are hallmarks of human cognitive priors. Recent studies show that video generation models, a paradigmatic class of current world models, have begun to show emerged reasoning abilities, making them ideal candidates for building human-like physical intelligence. Do video…
Just-in-Time Memory：LLMエージェント向けのタスク適応型メモリの学習
Agentic memory systems reuse past experience to improve future performance, yet most existing designs curate memory at write time: once a task is completed, its trajectory is distilled into a fixed artifact, such as a reflection, workflow, skill, or reasoning strategy, that is later retrieved by sim…
過去がフレームする未来：自己回帰型動画生成のためのメモリ
Advances in generative models have improved video fidelity, enabling long-horizon generation, interactive world modeling, and evolving visual environments. Autoregressive (AR) video generation extends visual sequences through causal rollouts. However, a fundamental bottleneck emerges: as the generat…
GrabとOpenAIが東南アジアに実践的なAIスキルをもたらす
OpenAI and Grab launch GO Forward with AI, a regional programme helping 30,000 partners build practical AI skills across Southeast Asia.
ClaudeがCRISPR様リピートを持つ新規酵素系を発見
In early results from our new life sciences research lab, Claude agents found an enzyme system whose function is still unknown.
GPT-6向けプロンプトキャッシング機能の改善
Learn how GPT-6 improves prompt caching with higher cache hit rates, new diagnostics, explicit breakpoints, and controls that reduce latency and costs.
GPT-6 Astra、Sol、Lunaを本番エージェント向けにMicrosoft Foundryで提供
Explore GPT-6 Astra, Sol, and GPT-6 Luna in Microsoft Foundry, with scalable model options for production AI agents, complex workflows, and high-volume tasks. The post GPT-6 Astra, Sol, and Luna: For production agents in Microsoft Foundry appeared first on Microsoft Azure Blog .
GPT-6 SolとLunaを発表
Meet GPT-6 Sol and Luna, two models that bring frontier intelligence to everyday work with different balances of capability and cost.
Claude Opus 5.5がMicrosoft Foundryで利用可能、長期実行コーディングと知識作業に対応
AI models are increasingly taking on work that extends far beyond a single prompt: building a feature across a codebase, investigating a complex issue, synthesizing hundreds of pages of information, or working through a multi-step business process. The post Claude Opus 5.5 comes to Microsoft Foundry…
Parallel、GPT-6 Astra により研究時間とコストを半減
GPT‑6 Astra allowed Parallel’s agents to research and synthesize labor-market data in half the time and at half the cost vs. prior models.
非集約量子化: LLMのプリフィルと復号の特化
Prefill and decode reward different approaches to quantization: low-precision arithmetic accelerates prompt processing, while compact weights reduce memory traffic during generation. We propose "disaggregated quantization" (DQ), which specializes computation formats, weights and storage placement to…
PACT：信用割当から批評家アライメントへ
Reinforcement learning has become a central component of large language model (LLM) post-training, yet token-level credit lacks a generally accepted mathematical definition, leaving its relationship to commonly used training signals unclear. We formulate three regularity conditions, namely Completen…
SpeakerMem-R1：マルチパーティー対話向けのスピーカー中心デュアルトラックメモリ
Long-term conversational memory in multi-party settings requires more than retrieving relevant content from long-term conversations: it must distinguish who said what, whom each statement concerns, how individuals perceive one another, what information is shared by the group, and how states change o…
The Tasteful Agent：長期タスクにおける美的感覚の測定と改善
LLM agents increasingly work on long-horizon tasks, and the decisions they make along the way, such as which hypothesis to test or which implementation to build on, determine the outcome of the whole run. Making these decisions well is becoming a key capability for both engineering and research agen…
Gemini 3.8 Flash TTS and Gemini 3.8 Flash-Lite TTS generally available (GA) : Released our next-generation text-to-speech (TTS) audio models and the Gemini API Voices endpoint ( /v1beta/voices ): Gemini 3.8 Flash TTS ( gemini-3.8-flash-tts ) : Flagship creative TTS model engineered for studio-grade…
Transformersがllama.cppの量子化に対応
UK AISIとEvalEvalがベンチマーク結果の再現性を実現
OpenAI APIチェンジログ 2026-09-22
Feature gpt-6-sol gpt-6-luna v1/responses v1/chat/completions Released GPT-6 Sol ( gpt-6-sol ) and GPT-6 Luna ( gpt-6-luna ). These reasoning models accept text and image inputs and generate text through the Responses and Chat Completions APIs. Standard pricing per 1M tokens for prompts with up to 2…
効果的なサードパーティ評価のための優先事項と原則
OpenAI outlines priorities and principles for rigorous, secure, and independent third-party AI safety assessments of frontier models and safeguards.
oMLXの開発者Jun KimがHugging Faceに参加、MLXコミュニティをサポート
OSWorld-Pro：コンピュータ操作エージェント向けプロセスベース評価
Evaluation of Computer-Use Agents (CUAs) is often limited to the final deliverables they create (at the end of hundreds of steps) and assessed with functional verifiers, as seen in OSWorld. However, such evaluation of end-state performance lacks transparency into how and why agents fail in various t…
物理学者のように LLM をプルーニング：Ising 最適化問題としてのブロック除去
Higgsfield AIがGPT-6 Astraで動画機能を1日で実装
With GPT-6 Astra, Higgsfield AI makes video ad creation easier for small businesses and brings new creative tools to market faster.
数学と人工知能に関する諮問委員会
OpenAI is working with an independent Advisory Group on Mathematics and Artificial Intelligence to guide the review and communication of emerging AI results.
次世代AI向けの標準化構築
OpenAI outlines a path to shared global AI standards, calling for coordinated evaluation, reporting, and governance to improve safety.
OpenAI Academyを新しいラーニングパスで拡大
Explore new OpenAI Academy learning paths for employees, developers, leaders, educators, and students to build and demonstrate practical AI skills.
量子増強拡散言語モデル向けCircuit Hypernetworks
Language models can be adapted by changing the computations applied to individual tokens. Quantum circuits offer one such approach, but evaluating wider circuits inside a large model can be computationally demanding. Here we introduce HyperQ, which adds token-conditioned quantum residual branches to…
Ovis-Embedding：汎用オムニモーダル埋め込みの最前線
In this report, we introduce Ovis-Embedding, a state-of-the-art omni-modal embedding family built on native integration of text, image, video, and audio. Instead of assembling separate modality towers, Ovis-Embedding uses a shared multimodal backbone to encode different modalities in a common repres…
HappyWorld-Bench
Evaluating world models requires assessing both the quality of the worlds they generate and their consistency and responsiveness under exploration, interaction, and modification. We introduce HappyWorld-Bench, a comprehensive benchmark that evaluates whether generated worlds remain reliable as agent…