What\u2019s new in AI tools
Automated weekly roundup of release notes and changelogs across every AI tool we track. Currently indexing 485 updates.
This week \u2014 2026-W18
AI Tools This Week: 42 updates across 7 products
-
Introducing Apache Spark 4.2
Apache Spark 4.2 has been released with enhancements that integrate more of the modern data and AI stack. The update brings new capabilities for data processing and AI workloads.
-
How Unity Catalog managed tables bring interoperability, performance, and unified governance to the Lakehouse
Unity Catalog managed tables enhance lakehouse architecture with improved interoperability and unified governance capabilities. The feature provides better performance and flexibility for enterprise data management at scale.
-
The skills gap behind agentic AI — and how Databricks is closing it with a new context engineer certification and agent trainings
Databricks launched a new Context Engineer certification program and agent training courses to address the skills gap in agentic AI development. The certification focuses on building AI agents and context engineering capabilities.
-
New calendar tools for your agent
Notion agents can now show and manage schedules, join calls, send invites, and find meeting times through chat. Available on desktop with mobile coming soon.
-
ChatGPT app, Dynamic voice speed, A/B testing, and more
Launched ChatGPT integration for building voice agents, dynamic voice speed matching caller pace, and A/B testing functionality for agent optimization.
-
Agent Versioning 2.0, Granular Multilingual, Custom Voicemail, Voice Orb, and more
Released comprehensive agent management system with version control, multi-language support with automatic detection, and enhanced voicemail handling capabilities.
-
Live Call Monitoring, Custom Dashboards, Built-in CRM, Colloquial Model, Expressive Mode, and more
Major platform update introducing real-time call monitoring, customizable dashboards, native CRM integration with Salesforce/HubSpot, and enhanced voice models for more natural conversations.
-
Groq Compound and Compound Mini
Production release of Groq's agentic AI systems that integrate web search, code execution, and browser automation into single API calls. Moved from beta to general availability with improved performance.
-
Moonshot AI Kimi K2 Instruct 0905
Added Kimi K2-0905 model with 256K context window, prompt caching support, and enhanced agentic coding capabilities. Features 200+ t/s performance at $1.50/M tokens blended pricing.
-
Remote Model Context Protocol (MCP)
Beta release of Remote Model Context Protocol server integration on GroqCloud, enabling AI models to connect to thousands of external tools through Anthropic's open MCP standard. Fully compatible with OpenAI APIs for seamless migration.
-
Introducing Mistral OCR 4
Mistral AI launched OCR 4, a state-of-the-art document intelligence model for optical character recognition tasks.
-
Bringing more control over your connectors
Enhanced connector functionality with improved control and management capabilities for enterprise integrations.
-
Leanstral 1.5: Proof Abundance for All
Mistral AI released Leanstral 1.5, an updated version of their proof-focused language model.
-
Your Prompts and Skills need a system of record
Studio now provides a system of record for AI prompts and skills with versioning, ownership tracking, and traceability features.
-
Introducing Robostral Navigate
Mistral AI released Robostral Navigate, their first model specifically designed for embodied navigation tasks.
-
GitHub Copilot CLI - Agentic coding agent for terminal
GitHub Copilot CLI allows developers to build, debug, and deploy code directly from the terminal with agentic AI capabilities. The tool includes GitHub MCP by default and supports extending with additional servers for enhanced context.
-
Agentic autofix for code scanning alerts in public preview
New automated code fixing capability for security alerts is now available in public preview.
-
Introducing Feature Views
Databricks introduces Feature Views to help ML teams build features once and reuse them across multiple models. This addresses the common problem of feature duplication in machine learning workflows.
-
Share Notion Workers across your team
Workers can now be shared with teammates with either 'Can connect' or 'Full access' permissions, enabling team collaboration on custom code solutions.
-
Ask Copilot for a repository overview
New feature allows users to request repository overviews directly from Copilot.
-
Share Notion Workers across your team
Users can now share Notion Workers with teammates using 'Can connect' or 'Full access' permissions. Enables team collaboration on Workers and cross-team agent deployment.
-
Share Notion Workers across your team
Users can now share Notion Workers with teammates, allowing collaboration on Worker development and connecting Workers to multiple Custom Agents across teams.
-
Share Notion Workers across your team
Teams can now share Notion Workers with different permission levels, allowing collaboration on AI automation tools across sales and marketing teams.
-
Notice Regarding Unauthorized Attempts to Sell Figure Stock
Figure AI issued a notice warning about unauthorized attempts to sell company stock. The company clarified its position on stock transactions and investor protections.
-
Meet the Notion Agents iOS app
Notion Agents are now available on iOS, allowing users to interact with their agents from mobile devices anywhere.
-
Meet the Notion Agents iOS app
Notion Agents are now available on iOS, allowing users to interact with their agents from anywhere on mobile devices.
-
Notice Regarding Unauthorized Attempts to Sell Figure Stock
Figure AI issued a notice warning about unauthorized attempts to sell company stock.
-
Meet the Notion Agents iOS app
Notion launched a dedicated iOS app for Agents, enabling users to interact with their AI agents from mobile devices anywhere.
-
Not All Miles are Equal: Why Time and Location Matter When Benchmarking Autonomous Safety
Waymo publishes research in Traffic Injury Prevention examining how time and location factors affect crash risk analysis. The studies highlight differences between highway commutes and downtown nightlife driving scenarios.
-
Not All Miles are Equal: Why Time and Location Matter When Benchmarking Autonomous Safety
Waymo published two peer-reviewed studies in Traffic Injury Prevention examining how time and location factors affect crash risk analysis for autonomous vehicles. The research addresses gaps in current safety benchmarking methods.
-
Not All Miles are Equal: Why Time and Location Matter When Benchmarking Autonomous Safety
Waymo publishes research on autonomous vehicle safety benchmarking methodology.
-
GitHub Copilot app available to all
GitHub Copilot app is now generally available to all users, expanding access beyond previous limitations.
-
Contextual Policies in Omnigent: Using session state to better govern AI agents
Databricks enhances Omnigent, their open source AI agent meta-harness, with contextual policies that use session state for improved governance. This allows for more sophisticated control over AI agent behavior.
-
Personal Injury Firm Cuts Discovery Drafting Time by 66% and Doubles Demand Output with AI Platform
Ostroff Godshall reduced discovery drafting time by 66% and doubled demand output while quadrupling monthly complaint filings using EvenUp's AI platform.
-
March Drop 2026
Integrated ChatGPT and Claude with company context via new remote MCP server and expanded third-party tool actions in Assistant and Agents.
-
May Drop 2026
Introduced real-time voice capabilities, company-aware Skills, and expanded action capabilities across existing team tools.
-
June Drop 2026
Added meeting notes functionality that turns meetings into searchable context with interactive content for team collaboration.
-
ChatGPT app, Dynamic voice speed, A/B testing, and more
Launched ChatGPT integration for building voice agents, dynamic voice speed matching caller pace, and A/B testing capabilities for comparing agent performance.
-
Agent Versioning 2.0, Granular Multilingual, Custom Voicemail, Voice Orb, and more
Released enhanced agent versioning with staging/production environments, granular multilingual support with automatic language detection, and custom voicemail handling capabilities.
-
Live Call Monitoring, Custom Dashboards, Built-in CRM, Colloquial Model, Expressive Mode, and more
Major platform update introducing real-time call monitoring, customizable dashboards, native CRM integration with Salesforce/HubSpot, and enhanced voice agent capabilities with colloquial phrasing and expressive delivery modes.
-
Moonshot AI Kimi K2 Instruct 0905 added
Kimi K2-0905 brings Moonshot AI's cutting-edge model to GroqCloud with enhanced agentic coding capabilities and a 256K context window. The model includes prompt caching for up to 50% cost savings and runs at 200+ tokens/second.
-
Remote Model Context Protocol (MCP) added
Remote Model Context Protocol (MCP) server integration is now available in Beta on GroqCloud, connecting AI models to thousands of external tools through Anthropic's open MCP standard. The implementation is fully compatible with OpenAI APIs, allowing developers to switch from OpenAI to Groq with zero code changes.
-
Custom instructions for personalized chat responses
GitHub Copilot now supports custom instructions in VS Code and Visual Studio to personalize chat responses based on preferred tools, organizational knowledge, and coding practices.
-
GitHub Copilot CLI - Agentic coding agent for terminal
GitHub Copilot CLI allows developers to build, debug, and deploy code directly from the terminal with AI assistance. The tool includes GitHub MCP by default and supports extending with additional servers for enhanced context.
-
Conductor, higher free-tier token limits, and more
Launched Conductor, an AI copilot for building and maintaining voice agents with simulation testing and automated change validation. Increased free LLM token allowance to 4,000 tokens per agent.
-
Leanstral 1.5: Proof Abundance for All
Mistral AI released Leanstral 1.5, an updated version of their open-source foundation model for trustworthy coding applications.
-
Notion 3.6: External Agents, HTML blocks, and more
Major release introducing External Agents (Claude, Cursor), interactive HTML blocks, speaker labels in AI Meeting Notes, Microsoft file support, and Outlook integration.
-
Notion 3.6: External Agents, HTML blocks, and more
Major release introducing External Agents integration with Claude and Cursor, interactive HTML blocks, speaker labels in meeting notes, and Microsoft Office file support.
-
Beyond dashboards: Introducing Decision Execution Platforms
Databricks announces Decision Execution Platforms, a new concept that extends beyond traditional dashboards to enable automated decision-making workflows.
-
How we keep GPUs reliable across Databricks AI
Databricks shares technical insights into their GPU reliability infrastructure and monitoring systems for distributed training workloads.