## OpenAI - **[Sora 2 Released:](https://openai.com/index/sora-2/)** OpenAI launched Sora 2, its advanced text-to-video generation model with improved video quality and generation capabilities. - **[OpenAI Launches Sora App:](https://techcrunch.com/2025/09/30/openai-is-launching-the-sora-app-its-own-tiktok-competitor-alongside-the-sora-2-model/)** OpenAI introduced a standalone Sora app functioning as a social network for AI-generated video content, positioning it as a TikTok competitor. - **[Parental Controls Added to ChatGPT:](https://www.engadget.com/ai/openai-is-adding-parental-controls-to-chatgpt-144128085.html)** OpenAI implemented parental controls in ChatGPT, allowing guardians to manage and monitor usage for younger users. - **[Instant CheckOut Launched:](https://chatgpt.com/merchants)** OpenAI introduced Instant CheckOut for merchants, enabling seamless AI-powered shopping experiences directly through ChatGPT. - **[Conversation Branching Feature:](https://x.com/OpenAI/status/1963697012014215181)** OpenAI introduced conversation branching in ChatGPT web, allowing users to explore different conversation directions without losing previous threads. - **[Developer Mode Launched:](https://platform.openai.com/docs/guides/developer-mode)** OpenAI unveiled Developer Mode for Plus and Pro users, providing full MCP access with enhanced tools management, connectors, and debugging capabilities. - **[Pulse Introduced:](https://openai.com/index/introducing-chatgpt-pulse/)** OpenAI launched Pulse, a $200/month Pro feature offering personalized daily AI-generated briefings. - **[Codex Updated with GPT-5:](https://openai.com/index/introducing-upgrades-to-codex/)** OpenAI updated Codex with GPT-5 capabilities, enabling longer autonomous coding sessions and improved code generation. - **[OpenAI-NVIDIA Partnership:](https://openai.com/index/openai-nvidia-systems-partnership/)** NVIDIA announced plans to invest up to $100 billion in OpenAI progressively to deploy 10 gigawatts of NVIDIA systems, with the first gigawatt deploying in the second half of 2026. ## Microsoft - **[VibeVoice-Large Released:](https://microsoft.github.io/VibeVoice/)** Microsoft released VibeVoice, a 10-billion-parameter model designed for long-form speech generation and voice synthesis. - **[Copilot Gets Avatars:](https://www.theverge.com/news/788741/microsoft-copilot-labs-ai-portraits-experiment-human-avatar)** Microsoft introduced avatar capabilities to Copilot through Copilot Labs, allowing users to interact with human-like AI portraits. - **[Anthropic Models in Copilot:](https://www.microsoft.com/en-us/microsoft-365/blog/2025/09/24/expanding-model-choice-in-microsoft-365-copilot/)** Microsoft integrated Anthropic's Claude models (Opus 4.1 and Sonnet 4) into Copilot, offering users more AI model flexibility. - **[New Microsoft Marketplace for AI Apps and Agents:](https://marketplace.microsoft.com/en-us/)** Microsoft unveiled a unified Marketplace as a hub to discover, try, buy, and deploy cloud solutions, AI apps, and agents. ## Google - **[Chrome Becomes Agentic Browser:](https://blog.google/products/chrome/new-ai-features-for-chrome/)** Google transformed Chrome into an agentic browser with Gemini built-in, enabling autonomous web navigation and task completion. - **[AI Mode Update for Visual Search and Exploration:](https://blog.google/products/search/search-ai-updates-september-2025/)** Google's AI Mode now supports visual search and conversational queries to help users explore topics more intuitively. - **[Gemini for Home Launched:](https://arstechnica.com/google/2025/10/googles-gemini-powered-smart-home-revamp-is-here-with-a-new-app-and-cameras/)** Google unveiled Gemini for Home, a comprehensive smart home platform powered by AI with a new app and camera integration. - **[Veo 3 Fast for YouTube Shorts:](https://blog.youtube/news-and-events/generative-ai-creation-tools-made-on-youtube-2025/)** YouTube introduced Veo 3 Fast for text-to-video generation, remixing, and AI editing tools specifically designed to simplify Shorts creation. - **[Mixboard App Released:](https://blog.google/technology/google-labs/mixboard/)** Google launched Mixboard, an AI-powered mood board application for design collaboration and creative brainstorming. - **[Learn Your Way Feature:](https://research.google/blog/learn-your-way-reimagining-textbooks-with-generative-ai/)** Google introduced "Learn Your Way," a tool that converts textbooks into interactive, personalized lessons. - **[AP2 Payment Standard:](https://cloud.google.com/blog/products/ai-machine-learning/announcing-agents-to-payments-ap2-protocol)** Google introduced AP2 (AI Payment Protocol 2) with Mastercard, PayPal, and American Express, enabling AI agents to securely handle cryptocurrency and multi-vendor transactions. - **[PayPal-Google Partnership:](https://newsroom.paypal-corp.com/)** PayPal partnered with Google to create AI-powered shopping experiences and support the new AP2 payment standard. - **[EmbeddingGemma Released:](https://developers.googleblog.com/en/introducing-embeddinggemma/)** Google launched EmbeddingGemma, a lightweight embedding model designed for on-device applications. ## Anthropic - **[Claude Sonnet 4.5 Released:](https://www.anthropic.com/news/claude-sonnet-4-5)** Anthropic launched Claude Sonnet 4.5, featuring improved performance and efficiency. - **[Claude for Chrome:](https://www.anthropic.com/news/claude-for-chrome)** Anthropic released Claude for Chrome, a browser extension allowing Claude to assist directly within the browser. - [**Claude Code Upgrades (VS Code Extension, Checkpoints)](https://www.anthropic.com/news/enabling-claude-code-to-work-more-autonomously):** Anthropic is enabling Claude Code to work more autonomously with a native VS Code extension, an enhanced terminal experience, and a checkpointing feature for long-running tasks. ## Meta - **[Meta Connect 2025:](https://www.meta.com/in/ai-glasses/meta-ray-ban-display/)** Meta held its annual Connect conference on September 17, unveiling Meta Ray-Ban Display smart glasses, Ray-Ban Meta Gen 2, and new Oakley Meta Vanguard sport glasses, along with developer features for Horizon OS. ## Other - [ **Grok 4 Fast Reasoning Model](https://x.ai/news/grok-4-fast):** xAI released Grok 4 Fast, a cost-efficient model excelling in agentic coding and quick reasoning tasks. - **[DeepSeek-V3.2-Exp Released:](https://api-docs.deepseek.com/news/news250929)** DeepSeek launched V3.2-Exp, an experimental version of their language model with enhanced capabilities. - **[Alibaba Qwen3-Next:](https://qwen.ai/blog?id=4074cca80393150c248e508aa62983f9cb7d27cd&from=research.latest-advancements-list)** Alibaba unveiled Qwen3-Next, a hybrid attention and sparse Mixture-of-Experts (MoE) LLM delivering 10x faster throughput with 90% lower compute cost. - **[Apple FastVLM and MobileCLIP2:](https://huggingface.co/collections/apple/fastvlm-68ac97b9cd5cacefdd04872e)** Apple released FastVLM and MobileCLIP2 models ahead of the iPhone 17 event, optimized for on-device vision-language tasks. - **[Periodic Labs AI Scientist Funding:](https://techcrunch.com/2025/09/30/former-openai-and-deepmind-researchers-raise-whopping-300m-seed-to-automate-science/)** Former OpenAI and DeepMind researchers raised $300 million in seed funding for Periodic Labs to build an AI Scientist for automating scientific research. - **[Notion AI Agent:](https://www.notion.com/blog/introducing-notion-3-0)** Notion unveiled its first AI agent capable of analyzing databases, generating reports, and automating workflows across Slack, Gmail, and Drive. - **[Zoom Companion 3.0:](https://news.zoom.com/ai-companion-3-0-and-zoom-workplace/)** Zoom launched Companion 3.0, introducing agentic AI capabilities for enhanced workplace productivity. - **[Gamma 3.0 Released:](https://siahualim.com/blog/technology/gamma-3-0-the-future-of-ai-powered-presentationsg)** Gamma 3.0 added an AI "co-pilot" for presentations and content creation. - **[Thinking Machines on LLM Consistency:](https://thinkingmachines.ai/blog/defeating-nondeterminism-in-llm-inference/)** Thinking Machines proposed solutions for addressing inconsistent outputs in large language models. - **[Atlassian Acquires The Browser Company:](https://www.atlassian.com/blog/announcements/atlassian-acquires-the-browser-company)** Atlassian acquired The Browser Company (makers of Arc and Dia browsers) for $610 million to build AI-first browsers. - **[Mistral Le Chat Updates:](https://mistral.ai/news/le-chat-mcp-connectors-memories)** Mistral AI updated Le Chat with MCP connectors and "Memories" feature for smarter workflow automation.