Who shipped what across the AI industry, tracked from official blogs and changelogs, the last 30 days.
The 12 busiest companies across the last 5 ISO weeks. Warmer cells shipped more; empty cells shipped nothing.
Launches per company over the window. A STREAK badge marks companies that shipped in every one of the last 3 or more ISO weeks.
What kind of work is shipping. One bar, 60 launches, split by category.
Every tracked launch, newest first. Filter by company, category, or search the text.
This blog post discusses the initiatives OpenAI is implementing to promote responsible AI development across Europe. It highlights policy engagement and collaborations aimed at ensuring ethical AI use in society.
The new Imagine Video 1.5 enhances video generation capabilities by incorporating text, image, and voice references, now allowing for native 1080p output. This update aims to improve motion, physics, and audio, with features like text-to-video and voice consistency becoming widely available across various platforms.
This blog post discusses the intricacies of autoscaling in the context of large language model (LLM) inference on the Together AI platform. It highlights the importance of selecting appropriate scaling metrics, tuning timing windows, and managing costs effectively while ensuring optimal performance under varying traffic conditions.
The web terminal in notebooks connected to serverless GPU compute (AI Runtime) is now in Public Preview. It allows users to run shell commands, monitor GPU usage, install libraries, and manage files directly on the compute environment.
This article provides continuous updates on the latest features and enhancements in Microsoft Fabric, showcasing new capabilities, improvements, and resources for users. Highlights include advanced functionalities, tools for development, and links to further information on various features.
The recent update introduces Projects, evolving from the previous concept of Spaces. Projects serve as collaborative hubs with a persistent file system and self-improving memory through the Computer Brain, supporting efficient teamwork and ongoing tasks.
Cursor has announced the formation of the Benchmark Partners Program, a global consortium of six strategic partners including AWS, BCG, Databricks, McKinsey, NVIDIA, and Snowflake. This initiative aims to assist enterprises in adopting AI effectively and transforming their software development processes.
Today marks the launch of Grok Voice Think Fast 2.0, our most advanced speech-to-speech voice model to date. Enhancements include improved intelligence, transcription accuracy, and conversational capabilities, setting a new benchmark in voice technology.
This post explains how Dedicated Model Inference on the Together AI platform operates, detailing the roles of endpoints, deployments, and configs, and how they interact through capacity-aware routing. The architecture allows for various features such as traffic splitting and A/B testing, providing insights into efficient model serving and scaling strategies.
This release notes document provides detailed information about the latest features and updates to the Adobe Experience Platform. Key highlights include new functionalities in Destinations, Real-Time CDP, and Segmentation Service that enhance data activation for cross-channel marketing and customer engagement.
This cumulative update for Windows 11 version 26H1 introduces numerous improvements, including a new calendar experience for update pausing, enhanced Widgets functionality, and accessibility options. Additionally, it addresses multiple reliability and performance issues across various system components, ensuring a smoother user experience.
Build Mode is a new feature from Grok that allows SuperGrok Heavy subscribers to create and share websites, apps, games, and interactive dashboards seamlessly in Early Beta. Users can describe their ideas, and Grok will generate a working version live for easy publishing.
Grok 4.5, the latest coding model, is now available in GitHub Copilot, enhancing the tools for millions of developers using VSCode and GitHub products. Developers can easily select the model from the Copilot settings or directly from the SpaceXAI console, which offers competitive pricing for usage.
Today, Perplexity launched Personal Computer for Windows, allowing users to access its agent platform on over a billion devices. This integration enables seamless workflow management across local files and Microsoft Office 365, enhancing productivity in enterprise settings.
The new Model Council feature in Computer lets users create a board of top AI models to tackle complex issues by choosing between models like OpenAI and Gemini. It synthesizes findings from selected models to produce actionable outputs, enhancing decision-making processes in various domains.
This post discusses the collaboration between Meta's AI models and the University of Pittsburgh's Human Engineering Research Laboratories to develop the Robotic Assistive Mobility and Manipulation Platform (RAMMP), aimed at enhancing mobility for people with disabilities. By integrating advanced robotics and AI technology, the initiative seeks to improve the lives of millions of wheelchair users in the U.S.
The NVIDIA Nemotron 3 Ultra, combined with the ACE-RTL agent, significantly enhances accuracy and efficiency in RTL workflows, achieving a 97.1% average pass rate across diverse tasks. This model optimally balances performance and cost, utilizing fewer tokens while outperforming leading models like GLM 5.2 and Kimi K2.6.
Today marks the launch of the V8.2 image model, which focuses on enhancing aesthetics and image quality. The update includes improvements in personalization, allowing for better understanding of user preferences and access to a broader selection of images.
Claude Opus 5 comes close to the frontier intelligence of Claude Fable 5 at half the price, and is priced the same as Opus 4.8 at $5 per million input tokens and $25 per million output tokens. It becomes the new default on Claude Max and the strongest option on Claude Pro.
Today we're bringing Grok into Google Workspace, allowing users to integrate Grok directly into Sheets, Slides, and Docs. This add-on enhances productivity by enabling users to ask questions and perform various tasks directly within their documents, such as writing formulas and creating presentations.
Runway Dev is a new AI media platform designed for developers, offering a single API for integrating top image, video, and character models. It aims to streamline the workflow for enterprise teams by providing robust security and tailored media solutions, already trusted by companies like Adobe and Figma.
A new health experience is being rolled out in ChatGPT, allowing eligible U.S. users to securely connect their health records and Apple Health data. This feature provides a centralized dashboard for viewing health information and asking related questions, while prioritizing privacy and security.
Grok Build now allows users to write and run workflows that can orchestrate complex tasks across hundreds of parallel agents, effectively verifying results in one background run. This feature is designed for tasks that require comprehensive reviews, such as large pull requests or extensive issue triaging.
The Galaxy Z Fold8 is a revolutionary foldable smartphone that embraces a compact and lightweight design while enhancing multitasking and content consumption with its advanced dual-display setup. It features top-tier performance powered by Snapdragon 8 Elite Gen 5 and boasts a robust camera system, making it the ultimate device for both productivity and entertainment.
At the Galaxy Unpacked event in July 2026, Samsung unveiled its eighth-generation Galaxy Z series, featuring advanced AI capabilities and three distinct foldable devices. The Galaxy Z Fold8 Ultra boasts impressive specifications with an ultra-thin display, while the Z Fold8 and Z Flip8 offer unique user experiences tailored to different lifestyles.
The Samsung Galaxy Unpacked event is set to take place in London on July 22nd, 2026, showcasing the latest devices including the Z Fold Ultra, Z Fold 8, and Galaxy Watch Ultra 2. Excitement is building, especially among fans eager to see new innovations and updates from Samsung.
Pre-order now and take advantage of exclusive offers valid from July 22, 2026, to August 14, 2026. Enjoy discounts and trade-in values on the latest Galaxy Z Fold8 Ultra, Z Fold8, and Z Flip8 devices.
Grok 4.5, our most intelligent model yet, is now available on grok.com, X, iOS, and Android. This update enhances the model's capabilities, providing smarter answers and the ability to manage tasks efficiently for various knowledge work.
On July 22, Samsung Electronics hosted Galaxy Unpacked in London, unveiling the newest additions to the Galaxy portfolio. The next generation of Galaxy devices combines intelligent capabilities with innovative form factors, delivering more personal and adaptive experiences while setting a new standard for the AI era.
The all-new Galaxy Z series opens a new chapter in foldables with three devices: the Z Fold8 Ultra, Z Fold8 and Z Flip8. The Z Fold8 pairs the Snapdragon 8 Elite Gen 5 Mobile Platform for Galaxy with dual 50 MP cameras, in a form factor that moves from the cover screen to the main display.
OpenAI and Hugging Face have collaborated to respond to a recent security incident, aiming to enhance security measures and protect user data. This initiative reflects their commitment to maintaining the integrity and trust in their platforms.
This post discusses the launch of the ChatGPT program tailored specifically for small businesses, aiming to empower them through AI technology. It highlights features, benefits, and the potential impact on small business efficiency and growth.
We’re introducing new Gemini models, including Gemini 3.6 Flash, 3.5 Flash-Lite, and 3.5 Flash Cyber, designed to enhance efficiency, latency, and reliability for AI agents at scale. The new models improve token efficiency, reduce costs, and build on customer feedback to support various developer workflows.
The Genesis Mission aims to accelerate scientific discovery using advanced AI, exemplified by the SYNAPS-I project that transforms data analysis in X-ray and neutron science. Leveraging Meta's open-source models, the initiative significantly reduces the time needed for image segmentation from weeks to around 15 minutes, enhancing research capabilities in various scientific fields.
The Grok add-in for Microsoft Outlook helps users summarize lengthy email threads, draft replies in their voice, and keep their inbox organized. It's designed to streamline communication and improve productivity when managing emails.
This post discusses the challenges and considerations regarding safety and alignment in advanced AI models that operate over long time horizons. It emphasizes the importance of ensuring these systems are developed responsibly to mitigate risks.
Anthropic is launching grant opportunities focused on rare genetic diseases as part of their AI for Science program, which aims to enhance scientific research and collaboration. Selected applicants can receive up to $50,000 in Claude credits to foster research into how AI can improve understanding and treatment of rare diseases.
Grok is now available as an add-in for Microsoft Excel, allowing users to ask questions in plain English, write formulas from descriptions, and run scenarios without leaving their workbook. The integration aims to enhance productivity by simplifying data interactions directly within Excel.
Today, we're introducing Automations: jobs Grok runs on its own. Describe the work once, choose when it runs, and Grok takes it from there, whether that’s research done before you're awake or an important email flagged the moment it lands.
Cohere and the University of Toronto have partnered to integrate Cohere’s AI technology into the university's platform. This collaboration aims to enhance responsible AI adoption, improving teaching, research, and administrative tasks while ensuring data security.
Hugging Face reported a significant security incident involving an AI-driven intrusion into its production infrastructure. The incident resulted in unauthorized access to internal datasets, but no public models were tampered with and major vulnerabilities were promptly addressed.
This post explores the concept of reliability in inference systems, detailing what different uptime percentages signify in terms of infrastructure and failure domains. It emphasizes the engineering challenges involved in maintaining high reliability and what potential clients should inquire about when choosing a provider.
The Microsoft 365 Copilot release notes detail the latest features and improvements, including changes in various applications like Word, Excel, and PowerPoint. Key updates highlight enhanced capabilities for AI-generated content, improved integration with existing workflows, and expanded options for users to personalize and streamline their interactions across Microsoft 365 services.
Perplexity introduces SPACE, a new sandbox platform designed to allow AI agents to operate securely while accessing necessary tools and files. This innovative architecture eliminates trade-offs between functionality, efficiency, and security by utilizing ephemeral sandboxes and robust security layers, ensuring sensitive information remains protected during long-running tasks.
SpaceXAI has opened the source code for Grok Build, its coding agent and terminal user interface (TUI). This release allows users to explore and extend the system, providing a definitive reference for its functionality and enabling fully local-first operation.
The July 2026 update of Visual Studio introduces a host of new features, including built-in .NET and Azure skills for improved developer productivity, enhanced GitHub Copilot functionality, and performance optimizations. This release marks a significant step towards integrating AI capabilities deeply into the development process, ensuring a smoother experience for developers.
Starting October 1, 2026, the specified runtime versions Python 2.7, Python 3.8, and PowerShell 7.1 & 7.2 will no longer be supported by Automation. It is strongly recommended to upgrade to supported runtime versions to minimize security risks and access the latest features.
This article explores the potential of Skills to enhance Codex's data analysis capabilities by turning implicit analytical conventions into explicit rules. Through experiments using the DataCOPE framework, it demonstrates that structured Skills can significantly improve accuracy in handling complex data analysis tasks, although they still require human validation.
GPT-5.6 represents a significant advancement in AI capabilities, enabling users to achieve more ambitious goals through enhanced intelligence. This new product aims to optimize performance and scalability, providing a robust tool for various applications.
Grok 4.5 is the latest flagship model offering enhanced capabilities for code and various tasks, featuring minimal hallucinations and configurable reasoning. It marks a significant upgrade aimed at providing intelligent and efficient solutions for diverse user needs.
This blog post addresses the public's hard questions about AI, including its potential effects on jobs, society, and medicine. Anthropic emphasizes the importance of understanding these concerns to ensure AI benefits society while mitigating risks.
Anthropic has launched a new beta feature designed for users to reflect on and refine their usage of Claude. This tool helps track and visualize AI interaction patterns, offering insights into when and how users engage with Claude, thereby supporting effective integration into daily life.
Most enterprises face challenges with unmanaged and scattered AI prompts and skills, leading to inconsistent behavior and issues that are difficult to trace. Mistral's Studio provides a centralized system for versioning, ownership, and traceability, allowing for rapid iteration and controlled deployment while ensuring compliance. By treating prompts as production assets, Studio enhances AI governance and discoverability, aligning AI behavior with business policies.
OpenAI has launched GPT-Live, a new generation of voice models designed to enhance natural human-AI interactions. This innovative technology allows for real-time conversation and integrates web search capabilities, making interactions feel more seamless and engaging.
The Claude apps gateway for AWS provides enterprises with a centralized control plane for managing access, cost, and policy for Claude Code and Claude Desktop. This self-hosted solution simplifies credential management, policy enforcement, and spend tracking, ensuring a secure and efficient development environment.
Hugging Face announces a new integration with Amazon SageMaker Studio, allowing developers to transition from model discovery to hands-on experimentation with just one click. This integration simplifies processes by pre-loading selected models and automating environment configurations, thereby enhancing the user experience for developers.
Cohere has announced the release of Cohere Transcribe Arabic, the most accurate open-source model for Arabic speech recognition, designed to address the dialectical diversity of the language. This model significantly outperforms its competitors, achieving the lowest word error rate among open-source models and enhancing transcription quality for both Arabic and bilingual Arabic-English audio.
Microsoft is eliminating around 4,800 roles, approximately 2.1% of its global workforce, to focus on priorities that will enhance its position in a fast-changing industry. The company aims to support those affected by these changes through redeployment efforts and by exploring alternative solutions to job eliminations.
Today, we're releasing 21 new flagship voices for Grok Voice, which are multilingual and designed for various use cases. The original five voices have also received improvements in naturalness, pacing, and emphasis, enhancing their delivery capabilities.
Cursor announced the CFO Council, a quarterly working group for finance leaders to develop frameworks for measuring AI productivity and managing costs. The post shares research finding that high token usage correlates with revenue growth, while AI benefits remain unevenly distributed across users and organizations.
Entries are pulled from official company blogs and changelogs only, no aggregators, no social posts. Each launch keeps its source link so every row is traceable.
Date sanity: any post that parses to a future date is treated as today, so a future date never appears on this page.
Data scraped AUG 02, 2026. Pulse refreshes weekly.