AI News & Updates
The latest from OpenAI, Anthropic, Google, xAI, and 100+ AI tools , updated live from RSS feeds.
Latest From the Big Four
The newest announcements from the leading AI models.
Anthropic Launches Claude Sonnet 4.6 and Opus 4.6
Claude Sonnet 4.6 is Anthropic's most capable mid-tier model yet, preferred over Opus 4.5 59% of the time in evaluations. Features a 1M token context window in beta, with major upgrades to coding, computer use, and agent planning.
Introducing GPT-5.2 for Professional Knowledge Work
GPT-5.2 is OpenAI's most capable model for complex multi-step tasks including coding, spreadsheets, and long-context reasoning. Sets a new state of the art on GDPval, outperforming industry professionals across 44 occupations.
Google Announces Gemini 3.1 Pro for Complex Reasoning
Gemini 3.1 Pro targets complex problem-solving with significantly higher benchmark scores. The February Gemini Drop also introduced Lyria 3 music generation, Veo 3.1 video templates, and Nano Banana 2 image generation.
xAI Launches Grok 4.20 With 4-Agent Collaboration
Grok 4.20 Beta introduces a 4 Agents collaboration system where specialized AI agents think in parallel and debate in real time before producing an answer. Available to SuperGrok and X Premium+ subscribers.
Showing 21 updates
Meta Introduces Llama 4 Scout and Maverick
Llama 4 Scout (17B active, 16 experts, 10M context) and Maverick (17B active, 128 experts) are natively multimodal with MoE architecture. Maverick matches DeepSeek V3 at less than half the active parameters.
GitHub Copilot CLI Reaches General Availability
GitHub Copilot CLI evolves from a terminal assistant into a full agentic development environment that plans, builds, reviews code, and maintains memory across sessions. Claude and OpenAI Codex now available as coding agents for Copilot Business.
Cursor 2.0 Launches Cloud Agents With Isolated VMs
Cursor 2.0 introduces Cloud Agents that run in isolated virtual machines, test their own changes, and produce merge-ready PRs. Accessible from desktop, web, mobile, Slack, and GitHub. Also ships Bugbot Autofix with 35% merge rate.
Notion 3.3 Introduces Custom AI Agents
Notion releases Custom Agents — fully autonomous AI that works around the clock without manual prompting. Assign a job, set a trigger, and they integrate with Slack, Figma, and more like a real teammate. Free to try through May 3.
DeepSeek V4 Targets Coding Dominance With Engram Memory
DeepSeek's upcoming V4 model uses Engram conditional memory for efficient retrieval from 1M+ token contexts. Internal benchmarks claim it outperforms Claude and GPT on long-context code generation.
Claude Code Gets Remote Control for Mobile Access
Anthropic releases Remote Control, a sync layer bridging local Claude Code CLI sessions with the mobile app and web interface. Start a complex task in your terminal and maintain control from a phone or tablet.
ElevenLabs Conversational AI 2.0 Goes Live
ElevenLabs launches Conversational AI 2.0 for enterprise-ready voice agents with major reliability improvements. Comes alongside a $500M Series D at ~$11B valuation and expanded Google Cloud GPU partnership.
Midjourney V8 Completes Final Rating Round Before Launch
Midjourney's V8 final rating party wraps up, signaling imminent release. V8 is a complete architectural rewrite on GPUs and PyTorch with native 2K resolution, improved text rendering, and advanced prompt understanding.
Kiro IDE Expands With Domain-Specific AI Powers
Amazon's Kiro IDE adds domain-specific "Powers" for HealthOmics bioinformatics, AWS Observability, and IAM Policy Autopilot. The spec-driven IDE requires structured specs before writing code, targeting enterprise quality.
Pinecone Introduces Dedicated Read Nodes for Scale
Pinecone launches Dedicated Read Nodes in public preview — provisioned capacity for predictable performance at scale. Supports 1000+ QPS, millions of namespaces per index, and high-performance sparse indexing.
Vapi Launches MCP Server for Voice Agent Integration
Vapi's new MCP server lets any MCP-compatible AI assistant control voice agents, manage calls, and phone numbers as native tools. Works with Claude Desktop and other LLM frameworks. Platform passes 350K developers.
Introducing n8n 2.0 With Publish/Save Workflow Paradigm
n8n 2.0 separates live production changes from draft edits with a new Publish/Save model, giving teams safer control over when updates go live. Laid the groundwork for Autosave and continued AI builder improvements.
Zapier Adds Draft/Published Versioning for AI Agents
Zapier ships version control for AI agents with one-click rollback, MCP Tool Bundle Sharing for importing all agent tools in a single link, and Guest Reviewers for external human-in-the-loop approval.
Kling AI 3.0 Generates Synchronized Video and Audio in One Pass
Kling 3.0 launches with a unified multimodal framework generating video and audio together, supporting native 4K at 60fps and multi-shot storyboards with per-shot camera control. Speech in 6 languages.
Supabase Ships PostgREST v14 With 20% Higher Throughput
Supabase's January update brings PostgREST v14 with ~20% higher request throughput. Also launched Supabase ETL (CDC to Apache Iceberg) in private alpha and AWS Marketplace purchasing for enterprise spend commitments.
Mistral Releases Devstral 2 and Vibe CLI
Devstral 2 (123B and 24B) achieves 72.2% on SWE-bench Verified, establishing itself as one of the best open-weight coding models. Vibe 2.0 launches as a terminal-native coding agent powered by it.
63% of Enterprises Have Fully Operationalized AI — Deloitte Report
Deloitte's 2026 enterprise AI report finds 63% of organizations have operationalized AI (up from 45% in 2025), workforce AI tool access grew 50%, and 85% expect to deploy autonomous agents. Only 21% have mature AI governance.
NotebookLM Integrates as a Source Inside the Gemini App
Google Workspace users can now add NotebookLM notebooks as a source in the Gemini app for grounded responses. Also gained chat-based slides, infographics, and podcast creation, powered by the Gemini 3 model.
FLUX.2 [klein] Generates Images in Under One Second
Black Forest Labs releases FLUX.2 [klein] in 4B and 9B variants — generating images in under one second on Nvidia GB200. Supports text-to-image, single-reference editing, and multi-reference composition. Apache 2.0 licensed.
Veo 3.1 Brings 4K Video at 60fps With Scene Extension
Google's Veo 3.1 adds 4K upscaling at 3840x2160 and 60fps, native 9:16 vertical video, and Scene Extension for continuous narratives exceeding 60 seconds. Generates native audio including dialogue, sound effects, and music.
Sora 2 Adds Reference Video and Dialogue Sync
OpenAI's Sora 2 generates 1080p video with improved physics, synchronized dialogue and sound effects, and the ability to insert specific people or objects from reference video. Now has a standalone iOS social app.
Never Miss an AI Update
Get the top AI news, new tool launches, and model updates delivered weekly, curated from 100+ sources.
No spam, unsubscribe anytime. Privacy Policy
Or find your perfect AI match
Take the Quiz