102 releases tracked across 20 AI companies in the past 30 days. Models, APIs, pricing, partnerships, acquisitions — all in one feed, refreshed weekly.
Today marks the launch of the Grok Build Plugin Marketplace, which features built-in plugins that enhance functionality within Grok Build. Users can browse, install, and update plugins directly from their terminal, making it easier to manage various tools and integrations.
Tori, eToro’s AI agent, is now enhanced with real-time market sentiment powered by SpaceXAI. This integration allows Tori to track live market moods and analyze signals, providing users with timely investment insights.
Gopuff and SpaceXAI have launched Go, an AI-powered shopping assistant integrated into the Gopuff app. This assistant utilizes advanced AI models to create personalized shopping experiences, ensuring quick deliveries based on user preferences and real-time data.
This announcement reveals a partnership between Grok and Vapi, making Grok the default engine for Vapi's core voices, significantly enhancing voice quality for over 2.5 million voice agents. The collaboration aims to offer more natural and emotionally resonant interactions for developers using the Vapi platform.
The latest image-to-video model, `grok-imagine-video-1.5-preview`, is now available via the xAI API in preview. This model enables the transformation of a still image into a cinematic video by utilizing natural-language prompts and maintaining detail from the original image.
Composer 2.5 is now available in Grok Build, offering a fast, state-of-the-art model designed for long-running tasks and complex instructions. Users can access it via the /models menu and it's available to SuperGrok and X Premium+ subscribers.
The latest updates and features for Grok Build have been documented in version 0.2.11, released on May 29, 2026. This version includes improvements such as a focusable tab bar, instant loading indicators, and various bug fixes to enhance user experience.
Grok Build 0.1, the fastest coding model from xAI, is now in public beta. Designed for agentic coding tasks like web development and debugging, it offers high performance at competitive pricing.
Starting today, users can integrate their Grok subscription directly within Kilo Code, an open-source coding platform designed for real software engineering workflows. This integration allows access to the latest Grok models without the need for a separate API key.
Transform a still image into a video by providing a source image along with your prompt. The model animates the image content based on your instructions, allowing for engaging video generation from a still frame.
This post introduces the integration of Grok into OpenCode, allowing users to utilize their SuperGrok or X Premium subscription. Users can easily set up OpenCode and connect to their Grok account to start coding with Grok Build, which powers xAI's terminal-based coding agent.
Starting today, users can utilize their SuperGrok or X Premium subscription within OpenClaw, an open-source, local-first agent and personal assistant designed to run on various hardware. OpenClaw allows seamless interaction with messaging platforms like WhatsApp and Telegram, providing persistent memory across sessions.
Today marks the launch of Skills, a new feature that allows Grok to remember user preferences across conversations. Users can create and manage custom skills without repetitive input, streamlining workflows and enhancing productivity across various formats.
xAI is retiring several earlier Grok models on May 15, 2026, to focus on the latest generation. From that date, requests to the retired models will be redirected to grok-4.3, which has a different pricing structure and reasoning efforts.
Starting today, you can integrate your Grok subscription with Nous Research’s open-source Hermes Agent, enhancing its capabilities for text conversations, text-to-speech, and image creation. This allows users to utilize the agent across messaging platforms while benefiting from its self-improving features and long-term memory.
The announcement details the launch of Claude Fable 5, a state-of-the-art model designed for general use, alongside Claude Mythos 5, which is tailored for select cyberdefenders with lifted safeguards. Both models demonstrate exceptional capabilities across various AI tasks, and while their launch aims to bring advanced AI capabilities to users promptly, safety measures are also emphasized to prevent misuse.
The Claude Partner Network introduces the Services Track and Partner Hub, designed to streamline partnership navigation in AI deployment. This initiative provides tiered structures for firms based on their Claude practices, facilitating integration and support for clients.
Anthropic is expanding Project Glasswing to include approximately 150 new organizations after initial deployment of Claude Mythos Preview. This collaborative effort aims to enhance software security and address vulnerabilities, focusing on critical infrastructure and industries like healthcare and communications.
Anthropic has secured $65 billion in Series H funding led by prominent investors, resulting in a post-money valuation of $965 billion. This funding aims to enhance safety research and increase the capacity to meet the growing demand for its AI systems, including Claude, which is being widely adopted across global enterprises.
Claude Opus 4.8 has been launched with significant improvements in efficiency and collaboration capabilities, enhancing user experience across various tasks. The model features dynamic workflows and allows greater control over task effort, all while maintaining the same pricing as its predecessor Opus 4.7.
Anthropic is expanding its presence in Europe with a new office in Milan, aimed at fostering collaboration with Italian enterprises and the developer community. This initiative coincides with significant discussions on AI ethics and its impact on society, led by Anthropic's commitment to responsible AI deployment in various sectors including finance, life sciences, and design.
Anthropic has appointed KiYoung Choi as the new Representative Director of Korea in anticipation of opening its Seoul office. With over three decades of experience in leading technology businesses, Choi aims to enhance the company's presence in a market that shows strong demand for AI solutions like Claude.
This post provides an update on Project Glasswing, a collaborative initiative focused on enhancing software security against AI-driven vulnerabilities. In its first month, the project has identified over ten thousand critical weaknesses across major software systems, demonstrating the urgency and necessity for improved cyber defense measures.
Claude Compliance API integrations with notable security and compliance tools have been introduced, allowing IT and security teams to govern Claude in the same manner as other applications in their stack. This feature enhances security and compliance across the Claude platform.
KPMG has announced a global alliance with Anthropic to integrate Claude into its business operations, granting access to over 276,000 employees. The partnership aims to innovate and improve client services in areas like tax, legal, and cybersecurity by embedding AI tools into KPMG's existing platforms.
At Anthropic, we are engaging diverse perspectives to address important questions raised by AI systems. Our discussions have included scholars and ethicists to inform the development of AI models like Claude, focusing on moral formation and the character of AI.
Anthropic is acquiring Stainless, a leader in SDKs and MCP server tooling, to enhance agent connectivity and developer experience. Founded in 2022, Stainless has been integral in powering Anthropic’s SDKs by converting API specifications into various programming language SDKs.
This blog post covers several new features and enhancements planned for Dynamics 365 Project Operations, focusing on efficiency and usability improvements such as a new mobile time entry app, enhanced contract management, and better invoice processing. These updates aim to streamline project management and enhance financial oversight for users.
At Microsoft Build 2026, the focus is on empowering developers amid a platform shift characterized by the duality of being a tinkerer and an enterprise builder. Key announcements include new tools and intelligence layers that enhance developer choice, security, and the ability to create bespoke agent systems, ensuring that developers can build effectively while maintaining governance and trust.
Microsoft Scout is the first Autopilot agent designed to work autonomously across Microsoft 365 apps, enhancing productivity by managing tasks and coordinating meetings proactively. This agent uses enterprise-grade security and learns user preferences over time, helping users stay organized and informed without constant prompting.
Microsoft Build 2026 highlights advancements in app development through Microsoft Fabric and Microsoft Databases, emphasizing a unified platform for scalable applications. This new approach addresses challenges in automation and data context, enabling developers to create production-ready agent systems efficiently.
This article outlines the key announcements for Microsoft Partner Center in June 2026, including updates on license keys for Dynamics 365 Business Central, Microsoft AI Skills Fest, and new prerequisites for Agent 365 licenses. It also covers the launch of the Work IQ API, updates on Solutions Partner skilling requirements, and the introduction of Azure Linux.
This blog post highlights the significant updates and improvements across the Fabric platform for June 2026, including enhanced data engineering capabilities, AI functions, and new event-driven features. Notable mentions include the upcoming community events in Barcelona and the introduction of advanced data management tools.
This release includes 162 Microsoft CVEs, detailing vulnerabilities and their respective exploitability ratings. Key updates cover a range of Microsoft products, with guidance on workarounds and mitigations for various threats.
This post outlines key announcements from Microsoft Partner Center for May 2026, highlighting significant developments such as the evolution of AI agents, expanded market opportunities for partners, and updates on various Microsoft programs and agreements. Key topics include new features in Microsoft 365, updates on partner agreements, and initiatives to boost partner marketing capabilities.
Microsoft's latest updates enhance security innovations across their ecosystem, focusing on AI-driven solutions for protecting data and identities. Enhancements include expanded visibility for Claude usage within Microsoft Purview and an advanced account recovery mechanism in Entra ID.
This article outlines the features planned for release in Dynamics 365 Sales from April 2026 to September 2026. It highlights upcoming functionalities related to sales productivity, lead management, opportunity acceleration, and sales engagement.
This blog post discusses the innovative ways in which an astrophysicist leverages OpenAI's Codex technology to create simulations of black holes. By integrating advanced AI tools, the astrophysicist enhances the accuracy and efficiency of these complex simulations, contributing significantly to the field of astrophysics.
The new feature enables web search to return image results along with regular text results, catering to applications that need current visuals. This expansion enhances user experience by providing relevant product photos and visual references.
OpenAI published a research post about improving ChatGPT memory. The item appears in the official research release listing on June 4, 2026.
OpenAI announced Codex-related product updates for a broader set of roles, tools, and workflows. The announcement appears in OpenAI’s official news and product release listings for June 3, 2026.
OpenAI announced that its frontier models and Codex are available on AWS. The post is listed under OpenAI Newsroom product releases and falls within the requested date window.
OpenAI has announced multiple updates across its models including the latest features and improvements to GPT-5.5 Instant, with enhanced response style and quality. The retirement of older models such as OpenAI o3 and GPT-4.5 is also highlighted, along with the introduction of new releases like GPT-5.4 mini. These changes aim to streamline user experience and improve conversational quality in ChatGPT and the API.
The latest release notes detail new features and updates for ChatGPT Enterprise and Edu users, including enhancements to Codex, improved integrations with Microsoft Teams and SharePoint, and feature introductions like Goal Mode and ChatGPT for Intune. These updates aim to enhance user experience and streamline workflows for educational and enterprise environments.
OpenAI's model has recently made a groundbreaking achievement by autonomously resolving the planar unit distance problem, a long-standing open question in discrete geometry. This milestone demonstrates the potential of AI in contributing original ideas and solutions to complex mathematical problems.
This blog post discusses the recent advancements in content provenance to ensure a safer and more transparent AI ecosystem. It highlights the importance of accountability in AI-generated content and offers insights into the measures being implemented.
OpenAI has launched a new personal finance experience within ChatGPT, enhancing the tool's capabilities to assist users with financial decisions. This release aims to provide tailored advice and insights for better money management.
The Google Colab Command-Line Interface (CLI) simplifies execution by connecting local terminals to Colab runtimes, enabling developers to provision powerful GPUs, execute scripts remotely, and recover artifacts easily. It is designed to integrate seamlessly with terminal environments, enhancing the workflow for both developers and AI agents.
Google's updates from May 2026 highlight a new 'agentic' era with the launch of the Gemini 3.5 model and Gemini Omni, enhancing tools for advanced reasoning and creation. The updates also include proactive applications like Universal Cart for shopping and the new Google Health app aimed at improving daily task management and personal wellness.
This blog post highlights the significant moments from Google I/O 2026, showcasing innovations like Gemini Omni and updates to Search. It covers various keynotes and product launches that promise to enhance user experience with AI-driven tools.
Google I/O 2026 featured numerous announcements and product launches, all compiled in a dedicated notebook using Google NotebookLM. The notebook includes various resources such as audio overviews, slide decks, highlights infographics, and more to keep users updated on key developments.
This blog post summarizes the major innovations and announcements made during Google I/O 2026, including the launch of new models like Gemini 3.5 Flash and Gemini Omni, enhancements in AI tools for search and shopping, and new features for the Gemini app. It highlights advancements in generative AI and how they are expected to improve user experiences across various platforms.
The Gemini API has launched support for managed agents, enabling developers to run custom agents in secure cloud sandboxes. Users can easily build and deploy these agents with their own instructions while leveraging the new Antigravity agent for enhanced functionality.
Google introduces significant AI enhancements to its Search, including an intelligent AI-powered Search box and new AI agents that provide personalized updates and assistance. These upgrades represent the most substantial changes to Search in over 25 years, aimed at making it easier for users to find and manage information using AI.
At Google I/O 2026, Google introduced new features for AI Studio, including native Android app development, Google Workspace integrations, and a mobile app for on-the-go development. These enhancements aim to streamline app creation and broaden the accessibility of AI Studio for developers of all levels.
At Google I/O 2026, we shared how we’re making AI more helpful for everyone, with the release of the Gemini Omni and Gemini 3.5 models. We are also enhancing our platforms with features like intelligent shopping carts and agent-first development environments, unlocking new possibilities across various Google products.
At Google I/O 2026, the focus is on enhancing productivity with Gemini, a new AI initiative designed to assist users proactively around the clock. The event showcases various features and innovative tools aimed at improving user experience and development capabilities.
Over nine months, our PR velocity has increased significantly as we've integrated Cursor Automations for security. This approach allows us to continuously identify and address vulnerabilities at scale, with four new automation templates designed for effective security monitoring and response.
We're excited to announce significant improvements to Bugbot, which is now over 3x faster, 22% cheaper, and finds 10% more bugs per review. With these advancements, 90% of Bugbot runs complete in under three minutes, facilitating quicker code merges and enhanced issue resolution.
Dropbox's CTO, Ali Dasdan, emphasizes the importance of quick AI adoption to maintain competitive speed. By integrating Cursor into its software development process, Dropbox has seen a significant increase in engineering velocity and tool adoption, with over 90% of engineers now using AI tools weekly.
Amplitude's engineering team leveraged Cursor to create a fully autonomous development pipeline, allowing for minimal developer intervention from idea to production. By using cloud agents, they streamlined processes like bug reporting and feature implementation, significantly boosting productivity and efficiency.
We are rolling out two improvements to our pricing based on user feedback. Users will now be able to pay for usage-based models with requests included in their subscription, and MAX will shift from per-tool-call pricing to API pricing, enhancing transparency and cost efficiency.
Gartner has recognized Cursor as a Leader in the 2026 Magic Quadrant for Enterprise AI Coding Agents, highlighting its position at the forefront of AI software development. This acknowledgment reflects the trust of over 70% of Fortune 500 companies in Cursor's innovative platform for managing coding agents throughout the software development lifecycle.
This blog post discusses the redesign of the Hugging Face CLI (`hf`) to better accommodate both human users and coding agents. It outlines improvements in output formatting, efficiency in handling complex tasks, and the introduction of features tailored for agent interactions, showcasing the CLI's performance compared to traditional methods.
This article discusses the advancements made with the DharmaOCR model, focusing on how Direct Preference Optimization (DPO) serves as a solution for text degeneration in OCR. It presents empirical evidence showing that DPO effectively reduces error rates in various model families, emphasizing the need to address binary correctness rather than simply maximizing likelihood in training.
This article presents an experiment comparing shallow and deep transformer models to explore the impact of model depth on reasoning. The study shows that a 12-layer model significantly reduced training loss, providing evidence that deeper models can better fit reasoning patterns despite a moderate increase in parameters.
This post introduces the importance of profiling in optimizing PyTorch applications, noting its steep learning curve. Through a series of blog posts, it guides readers in understanding how to read and utilize profiler traces for better performance in machine learning workflows.
This blog post introduces the trimming technique, a method used for creating lighter models without requiring retraining, while maintaining performance. It explores the benefits of this technique for multilingual models, demonstrating its efficacy through collaborative efforts of various Hugging Face Fellows.
This post discusses the latest updates, new features, and bug fixes in the recent Illustrator on desktop releases. It covers changes in versions May 2026, April 2026, March 2026, and others, highlighting key functionalities and issues that have been resolved, as well as new features introduced to enhance user experience.
This month's update introduces an on-device Remove Tool powered by a downloadable AI model, improvements to Live Filter Layers for non-destructive filtering, and the ability to send documents directly to Firefly Boards. A range of bugs related to generative AI and file handling have also been resolved, enhancing user experience significantly.
This blog post outlines various enhancements made in the Third Quarter 2026 release of the Preview environment. Key updates include a redesigned email format for comment notifications, integration with AI tools through the Workfront MCP Server, and improved features for enhanced lists.
Photoshop 27.7 introduces enhanced performance and AI capabilities, making it easier for users to focus on creativity rather than complexity. This update streamlines workflows, improves efficiency, and includes features like an on-device AI model for the Remove tool and direct integration with Firefly Boards.
This blog post covers the latest updates and features introduced in Adobe Photoshop with the recent versions, including enhancements in AI capabilities, new integrations, and performance improvements. It outlines key feature updates, enhancements, and several customer-reported fixes as of May 20, 2026.
At the AI Now Summit, Mistral announced an integrated AI stack aimed at transforming industrial operations through partnerships with Airbus, BMW Group, and ASML. Key initiatives include leveraging advanced physics AI to optimize designs and improve operational efficiencies, as well as the unveiling of the new Vibe agent for enhanced productivity.
Mistral introduces Vibe, a powerful AI agent designed for managing complex, multi-step tasks across various workflows, including coding and task orchestration. It features both Work and Code Modes, allowing users to automate their processes efficiently while integrating seamlessly with existing tools and platforms.
Mistral's acquisition of Emmi AI underscores its dedication to advancing AI research and enterprise solutions in industrial engineering. This initiative aims to enhance product development speed while ensuring operational efficiency in key sectors like aerospace and automotive.
Mistral is unveiling its new physics AI capabilities, aimed at revolutionizing the engineering sector by enhancing the speed and quality of design processes. This technology allows for faster simulations and real-time insights, enabling companies to explore thousands of design variants efficiently and improve overall product performance.
Mistral AI has announced its acquisition of Emmi AI to enhance its capabilities in industrial AI and physics modeling. This strategic move aims to position Mistral as a leader in providing innovative solutions for the engineering domain, utilizing advanced AI technologies to revolutionize engineering workflows.
The Amazon Managed Service for Prometheus has introduced support for out-of-order sample ingestion and a workspace-level rule query offset. These features aim to reduce data loss and enhance alerting accuracy for workloads affected by variable network latency and delayed data arrival.
AWS Security Agent now generates verification scripts for penetration test findings, allowing security teams to easily reproduce and validate vulnerabilities. This automatic generation of ready-to-run scripts simplifies the process, enhancing triage and speeding up remediation efforts.
AWS Security Hub Extended has introduced 21 curated partner solutions spanning 9 security categories, enhancing enterprise security flexibility. This expansion includes solutions from notable partners like SentinelOne and CyberArk, ensuring a more comprehensive security strategy.
AWS has launched the public preview of AWS Interconnect, enabling multicloud connectivity with Oracle Cloud Infrastructure (OCI). This innovative product allows for the seamless provisioning of resilient, scalable private connections to other cloud providers, streamlining the interoperability of applications across multiple cloud environments.
Runway's fourth annual AI Festival celebrates the intersection of art and technology. Showcasing AI films and creative works, it highlights how today's creatives are using advanced tools to reshape artistic expression across various fields.
Runway has launched Runway MCP (Model Context Protocol), an MCP server that integrates with existing agents and coding tools, enabling users to generate images and videos directly within their workflow. This tool allows for the seamless creation of marketing assets and other visual content without switching between applications, boosting productivity for creative projects.
The article discusses the advancement of AI-generated video technology, highlighting its ability to create narratives that engage audiences emotionally rather than just showcasing technical prowess. It describes a shift where viewers are now able to appreciate stories rather than focus on imperfections in the AI media, indicating a maturation in the medium.
Aleph 2.0 is the latest upgrade to Runway's flagship video editing model, launched alongside Edit Studio. This new product offers enhanced video editing capabilities, allowing for targeted changes while preserving original content, and is designed to help users create the video they need with greater efficiency and control.
This blog post details the new features and improvements released in June 2026 for the Databricks platform. Highlights include enhancements to the Managed SharePoint connector, improvements in workspace file performance, and updates to Genie products.
Databricks has unveiled the keynote lineup for the 2026 Data + AI Summit, highlighting its status as the world’s largest gathering focused on data and AI. The conference promises an extensive program featuring industry leaders and innovative sessions.
The new query-based connectors in Lakeflow Connect allow data ingestion from various databases by querying the source directly. This method eliminates the need for change data capture (CDC) configuration or an ingestion gateway, supporting sources like Oracle, SQL Server, and PostgreSQL.
This update outlines the new features, improvements, and bug fixes for Lakeflow Spark Declarative Pipelines released in 2026. Significant enhancements include support for serverless compute, advanced data processing capabilities, and dynamic resource adjustments, among others.
Perplexity announces the Main Street AI Accelerator, a $25 million initiative in collaboration with the U.S. Small Business Administration, aimed at supporting growing businesses with AI tools. The program offers $250 in credits to up to 100,000 eligible companies, facilitating access to essential resources for business growth.
Perplexity has announced the launch of Personal Computer for Windows, enhancing the computing experience by allowing users to interact with local files and Microsoft apps seamlessly. This integration aims to improve productivity by bridging the gap between local and web-based tasks.
The Samsung Galaxy S25 series is now receiving exciting new AI features through its June 2026 software update. This update introduces advanced tools for notification and document management, enhancing productivity by summarizing notifications and generating concise text summaries of files.
The recent update to the A16 5G phone has resulted in notable changes, including the removal of sub-header section names in the Settings menu. Users now have to rely on memory or the search function to locate settings, raising concerns about usability and convenience.
The Groq Community is transitioning, with plans to retire the community site on May 29, 2026. Although the platform will remain active, new posts will be disabled starting May 8, and existing content will remain accessible.
The Groq Community will be sunsetting on May 29, 2026, although the platform itself will remain available. New posts will be disabled beginning May 8, but existing content will remain accessible until the site goes offline.
Cohere celebrates key milestones with branded merchandise that showcases its culture and spirit. The newly launched Cohere Merch Store offers fans and customers the chance to own a piece of Cohere's history, marking the evolution of its aesthetic over time.
Cohere has launched Command A+, an advanced open-source language model designed for critical infrastructure. With a focus on data sovereignty and efficiency, it supports 48 languages and aims to empower enterprises and public sectors with secure AI deployment.
Improvements to conversational mode for voice and text inputs, making it easier to access image prompts and recent jobs. A new 'Rerun as HD' feature is introduced, along with several bug fixes and enhancements to the mobile web experience.
This blog post discusses NVIDIA's Medical AI for Synthetic Imaging (MAISI) initiative, which aims to synthesize high-resolution 3D medical imaging data to overcome challenges of data scarcity and privacy restrictions in training AI models. The highlight is the introduction of NV-Generate-MR-Brain, a generative model for creating realistic brain MRI and segmentation data, enhancing the capacity for medical AI development.
The release of Stable Audio 3.0 introduces a new model family designed for generative audio, featuring open-weight music models trained on fully licensed data. Key advancements include variable-length audio generation and full song composition capabilities on portable devices, enhancing creative experimentation in the audio community.
This post discusses the efficiency of the Together Inference Engine in handling coding agent workloads, highlighting its performance improvements over competitors like TensorRT-LLM and SGLang. The benchmarking results show Together achieving 31% more TPS, with better performance under concurrent user loads and latency conditions.