## **July 2025** ## Google - **[NotebookLM Video Overviews and an upgraded Studio](https://blog.google/technology/google-labs/notebooklm-video-overviews-studio-upgrades/**): NotebookLM can now generate Video Overviews to distill complex information into clear, digestible content, paired with a powerful, redesigned Studio panel. - **[OPAL No-Code Agent:](https://developers.googleblog.com/en/introducing-opal/)** Google introduced **OPAL** (Objective-driven Prototyping and Application Launcher), a no-code agent designed for building business applications. Users can describe their desired app in natural language, and the agent autonomously creates a fully functional prototype. - **[Firebase Studio Adds Agent Mode:](https://firebase.google.com/docs/studio)** Google's **Firebase Studio** now includes an agent mode, allowing developers to use a Gemini-powered agent to assist with prototyping, debugging, and code generation for full-stack applications. - **[Gemini Code Assist Agent Mode:](https://developers.google.com/gemini-code-assist/docs/use-agentic-chat-pair-programmer)** **Gemini Code Assist** for developers now has an "agent mode" that gives it full remote codebase awareness, enabling it to execute multi-file changes and other complex tasks from a single prompt. - **[Veo 3 Video Model Released:](https://deepmind.google/models/veo/)** Google rolled out its video generation model, **Veo 3**, which can create 8-second video clips from text prompts. - **[Gemini on Pixel Watch:](https://blog.google/products/pixel/pixel-watch-4/)** The **Gemini** assistant is now available on the new Pixel Watch, replacing the older Google Assistant and offering more advanced conversational AI features on the wearable device. ## Microsoft - **[Deep Research in Azure AI Foundry:](https://www.google.com/search?q=https://www.microsoft.com/en-us/research/blog/deep-research-azure-ai-foundry/)** Microsoft has introduced **deep research** capabilities in its Azure AI Foundry, which enable the creation of highly specialized AI agents for scientific and academic research. These agents can autonomously explore and synthesize large bodies of text to form new hypotheses. - [**GitHub Vibe Coding with Spark](https://github.blog/changelog/2025-07-23-github-spark-in-public-preview-for-copilot-pro-subscribers/):** Microsoft launched **GitHub Spark**, an AI-powered micro-app builder. Spark allows developers to describe an application in plain English and instantly generate a deployable mini-app, part of a new "vibe coding" movement. ## OpenAI - [**Introducing study mode**](https://openai.com/index/chatgpt-study-mode/): OpenAI launches Study Mode, a new way to learn in ChatGPT that offers step-by-step guidance instead of quick answers. - [**Introducing ChatGPT agent: bridging research and action**](https://openai.com/index/introducing-chatgpt-agent/): OpenAI introduces ChatGPT Agent mode, allowing ChatGPT to complete complex tasks through tools like a visual browser. - [**Introduction to deep research in the OpenAI API**](https://cookbook.openai.com/examples/deep_research_api/introduction_to_deep_research_api): OpenAI releases the deep research API, enabling automation of complex research workflows that require reasoning, planning, and synthesis. - [**OpenAI and Gemeni achieve gold medal standard at the International Mathematical Olympiad**](https://deepmind.google/discover/blog/advanced-version-of-gemini-with-deep-think-officially-achieves-gold-medal-standard-at-the-international-mathematical-olympiad/): OpenAI and Google's Gemini achieve gold medal level performance at the 2025 International Mathematical Olympiad, solving complex problems. ## Anthropic - **[Claude for Financial Services:](https://www.anthropic.com/news/claude-for-financial-services)** Anthropic announced a comprehensive solution, **Claude for Financial Services**, which leverages its models to accelerate financial analysis and compliance tasks. - **[Free Academy Courses:](https://anthropic.skilljar.com/)** The company launched **Free Academy Courses** to help developers and enterprises learn how to build and deploy applications using Anthropic's models and tools, including the Model Context Protocol. ## Others - [**Voxtral**](https://mistral.ai/news/voxtral): Mistral open-sources Voxtral, a frontier open-source speech understanding voice model. - [**Upgrading agentic coding capabilities with the new Devstral models**](https://mistral.ai/news/devstral-2507): Mistral introduces Devstral 2507, including Medium and Small upgrades for agentic coding. - [**Announcing Codestral 25.08 and the Complete Mistral Coding Stack for Enterprise**](https://mistral.ai/news/codestral-25-08): Mistral releases Mistral Code and Codestral 25.08 for enterprise-grade software development. - [**Alibaba Unveils Cutting-Edge AI Coding Model Qwen3-Coder**](https://www.alibabacloud.com/press-room/alibaba-unveils-cutting-edge-ai-coding-model-qwen3): Alibaba releases Qwen Coder, an AI coding model with agentic capabilities. - [**Enabling customers to deliver production-ready AI agents at scale**](https://aws.amazon.com/blogs/machine-learning/enabling-customers-to-deliver-production-ready-ai-agents-at-scale/): Amazon AWS introduces Kiro Agentic IDE for spec-driven development from concept to production. - [**Amazon Bedrock AgentCore now available in preview**](https://aws.amazon.com/about-aws/whats-new/2025/07/amazon-bedrock-agentcore-preview/): Amazon AWS launches Bedrock AgentCore for deploying and operating AI agents securely at scale. - [**Enabling customers to deliver production-ready AI agents at scale**](https://aws.amazon.com/blogs/machine-learning/enabling-customers-to-deliver-production-ready-ai-agents-at-scale/): Amazon AWS plans to launch an AI Agent Marketplace, offering a centralized catalog of curated agents and tools. - [**Cursor on Web and Mobile**](https://cursor.com/blog/agent-web): Cursor expands to mobile and web, allowing work with powerful coding assistants anywhere. - [**Open Deep Research**](https://blog.langchain.com/open-deep-research/): Langchain open-sources their open deep research tool, configurable with custom models and search tools. - [**Reachy Mini - The Open-Source Robot for Today's and Tomorrow's AI Builders**](https://huggingface.co/blog/reachy-mini): Huggingface releases Reachy Mini, an expressive open-source robot for AI experimentation. - [**We Built an Operating System for AI...But Is It Really One?**](https://medium.com/%40warmwind/we-built-an-operating-system-for-ai-but-is-it-really-one-e4d9ef3fec97): Warmwind launches Warmwind OS, an agentic OS for automating digital workflows with AI agents. - [**Introducing cloud browser**](https://manus.im/help/cloud-browser): Manus introduces Cloud Browser, an agentic browser that syncs login states and performs tasks autonomously. - [**Cognition's acquisition of Windsurf**](https://cognition.ai/blog/windsurf): Cognition acquires Windsurf, the agentic IDE, including IP and product. - [**NVIDIA Canary‑Qwen‑2.5B: Open‑Source ASR/LLM for Superior Transcription and Summarization**](https://forums.developer.nvidia.com/t/nvidia-canary-qwen-2-5b-open-source-asr-llm-for-superior-transcription-and-summarization/339387): Nvidia introduces Canary-Qwen-2.5B, an open-source hybrid ASR and LLM model. - [**Lovable Raises $200M, Valued at $1.8B, Just Eight Months After Launch**](https://lovable.dev/blog/200m-series-a-fundraise): Lovable raises $200 million in funding. - [**From Googlebot to GPTBot: who's crawling your site in 2025**](https://blog.cloudflare.com/from-googlebot-to-gptbot-whos-crawling-your-site-in-2025/): Cloudflare blocks AI Crawlers by default unless compensated, introducing pay-per-crawl mechanisms.