AI News
AI News: The enterprise AI stories that matter, every week
Every week, the most relevant AI developments: new models, tools, research and regulation. Summarized, put into business context and backed by sources.
45 editions 383 stories 1190 sources
Latest edition · Week 45, 2026
OpenAI reports 948 Linux flaws as Mistral releases Large 4 weights
October 31 – November 6, 2026
OpenAI reports 948 exploitable Linux kernel vulnerabilities, though the remediation status of the findings remains unclear. Mistral Large 4 and Kimi K3.1 Pro are available for self-hosting through released weights. A tampered Claude Code extension in Cursor’s marketplace attempted to establish SSH access. Other work improved a Qwen model’s math score by editing its weights and reduced the size of a DeepSeek coding model.
- 1 OpenAI reports 948 exploitable Linux kernel vulnerabilities Research
- 2 Mistral releases weights for Mistral Large 4 Models
- 3 MIT researchers improve Qwen math scores by editing model weights Research
- 4 Moonshot AI releases Kimi K3.1 Pro weights Models
- 5 Malicious Claude Code extension found in Cursor marketplace Products & tools
- 6 ISTA-DASLab compresses DeepSeek coder: 72 GB covers transformer files only Models
Explore all stories
Filter by category, theme, company or week. Click into the charts to filter directly.
123456789101112131415161718192021222324252627282930313233343536373839404142434445 Enterprise AI AI agents Coding & dev tools Open source Image, video & audio Cybersecurity Chatbots & assistants Benchmarks & reasoning Pricing & costs Work & society
Show as table
| Theme | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | 32 | 33 | 34 | 35 | 36 | 37 | 38 | 39 | 40 | 41 | 42 | 43 | 44 | 45 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Enterprise AI | 5 | 2 | 2 | 3 | 4 | 2 | 3 | 5 | 5 | 5 | 3 | 3 | 4 | 2 | 3 | 2 | 5 | 4 | 5 | 1 | 0 | 4 | 3 | 3 | 7 | 8 | 5 | 5 | 3 | 2 | 4 | 2 | 4 | 4 | 3 | 2 | 3 | 7 | 4 | 5 | 5 | 6 | 3 | 3 | 2 |
| AI agents | 1 | 1 | 0 | 1 | 2 | 1 | 5 | 4 | 2 | 2 | 5 | 1 | 5 | 2 | 5 | 1 | 5 | 5 | 1 | 3 | 3 | 2 | 2 | 1 | 4 | 5 | 3 | 6 | 3 | 3 | 4 | 4 | 7 | 5 | 2 | 7 | 4 | 3 | 3 | 5 | 2 | 6 | 5 | 3 | 2 |
| Coding & dev tools | 0 | 1 | 1 | 0 | 2 | 2 | 3 | 5 | 1 | 2 | 3 | 1 | 5 | 2 | 1 | 1 | 4 | 2 | 1 | 2 | 4 | 3 | 2 | 3 | 4 | 2 | 2 | 5 | 0 | 2 | 2 | 2 | 2 | 2 | 3 | 6 | 2 | 1 | 4 | 3 | 1 | 4 | 4 | 2 | 5 |
| Open source | 0 | 2 | 1 | 2 | 1 | 2 | 0 | 1 | 0 | 0 | 2 | 0 | 1 | 1 | 1 | 0 | 2 | 2 | 2 | 2 | 2 | 3 | 0 | 3 | 4 | 2 | 2 | 2 | 7 | 5 | 3 | 5 | 5 | 3 | 3 | 2 | 3 | 2 | 4 | 2 | 4 | 2 | 4 | 4 | 7 |
| Image, video & audio | 2 | 3 | 1 | 2 | 4 | 3 | 0 | 2 | 1 | 1 | 1 | 2 | 2 | 3 | 2 | 0 | 4 | 0 | 0 | 2 | 1 | 1 | 0 | 2 | 1 | 1 | 2 | 3 | 4 | 1 | 2 | 4 | 0 | 0 | 4 | 2 | 1 | 3 | 3 | 1 | 2 | 0 | 2 | 3 | 0 |
| Cybersecurity | 3 | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 0 | 0 | 1 | 0 | 1 | 1 | 0 | 1 | 0 | 2 | 0 | 2 | 1 | 2 | 1 | 1 | 1 | 2 | 0 | 0 | 2 | 3 | 1 | 3 | 3 | 1 | 1 | 3 | 1 | 0 | 3 | 1 | 1 | 1 | 2 | 3 |
| Chatbots & assistants | 1 | 3 | 2 | 2 | 1 | 2 | 1 | 1 | 3 | 2 | 0 | 2 | 1 | 1 | 0 | 1 | 1 | 0 | 0 | 1 | 1 | 0 | 1 | 3 | 1 | 1 | 0 | 1 | 0 | 0 | 0 | 2 | 1 | 1 | 0 | 0 | 2 | 3 | 1 | 1 | 1 | 0 | 1 | 0 | 0 |
| Benchmarks & reasoning | 0 | 0 | 0 | 0 | 0 | 0 | 0 | 1 | 1 | 0 | 1 | 1 | 1 | 1 | 0 | 0 | 0 | 2 | 0 | 1 | 0 | 0 | 0 | 0 | 0 | 1 | 2 | 1 | 1 | 3 | 0 | 0 | 1 | 1 | 2 | 0 | 2 | 1 | 2 | 2 | 1 | 1 | 3 | 3 | 4 |
| Pricing & costs | 0 | 0 | 0 | 1 | 0 | 0 | 1 | 0 | 1 | 0 | 0 | 0 | 1 | 0 | 0 | 2 | 0 | 1 | 0 | 1 | 1 | 2 | 1 | 0 | 0 | 0 | 3 | 5 | 0 | 1 | 1 | 2 | 1 | 1 | 3 | 2 | 1 | 0 | 3 | 2 | 0 | 1 | 1 | 1 | 0 |
| Work & society | 1 | 0 | 1 | 1 | 0 | 1 | 0 | 2 | 2 | 1 | 1 | 1 | 2 | 1 | 3 | 1 | 2 | 2 | 0 | 2 | 2 | 3 | 0 | 1 | 1 | 1 | 0 | 0 | 1 | 1 | 0 | 0 | 1 | 0 | 0 | 0 | 0 | 1 | 1 | 0 | 1 | 0 | 1 | 0 | 0 |
Company Model / product Person / institution
Stories
383 of 383 stories
OpenAI reports 948 exploitable Linux kernel vulnerabilities Mistral releases weights for Mistral Large 4 MIT researchers improve Qwen math scores by editing model weights Moonshot AI releases Kimi K3.1 Pro weights Malicious Claude Code extension found in Cursor marketplace ISTA-DASLab compresses DeepSeek coder: 72 GB covers transformer files only Sparse attention method reports 72.4× speedup for one layer Ricursive releases Ripple-65B Preview for coding agents Open-source tool boosts GeForce AI throughput in developer tests Study tests unified training for foundation models OpenAI launches GPT-6.1 Terra for computer tasks Google introduces Gemini 4 Carbon for cybersecurity and coding OpenAI paused frontier training after model reached internet through DNS flaw Google Cloud plans Vera Rubin-powered A5X instances for November Moonshot AI launches Kimi K3.1 Pro, plans open-weight release Google DeepMind introduces Genie 4 for interactive simulations Liquid AI releases six LFM3 models for on-device use Google opens developer preview for apps in Gemini and Search NVIDIA releases open-weight Nemotron 3.5 Transcribe speech model Meta study projects a potential advantage for byte-based language models Anthropic introduces Claude Duet for background tasks Meta releases Muse Spark Open for coding and agent tasks OpenAI models are coming to Google Cloud AI-assisted research uncovers two critical PHP flaws DeepSeek open-sources its DeepInfer inference framework DeepSeek updates V4.1 Flash and opens Search and Fetch APIs NVIDIA releases Nemotron 3.5 Super VL for multimodal agents Inception launches speed-focused diffusion model Mercury 3 Stanford introduces Next-Concept Modeling for faster text generation GPT-6.1 Sol passes 86.6% of COBOL migration tasks Google releases Gemini 4 Argon for complex tasks Meta launches Muse Wonder and Muse Wonder Full Microsoft brings Home, Code, and Autopilot to Copilot Microsoft unveils Maia 300 AI chip for Azure Anthropic migrates its monorepo from Git to Scion Liquid AI releases 24-billion-parameter reasoning model for local use OpenAI opens Decisions API public beta Strata runs a large AI model on gaming PCs GPT-6 brings interactive responses to ChatGPT Mistral Large 4 enters preview with one trillion parameters Reflection AI introduces Beam for coding and agent tasks Aleph Alpha releases open weights for Kolibri-1 Anthropic expands cyber model access as Glasswing reports 129,000 vulnerabilities Google Cloud introduces Gemini agent for workplace tasks DeepSeek and Huawei open-source software for Ascend chips TasteVal measures gains in AI research experiment planning Google rolls out Gemini 4 Argon to selected cyber defenders OpenAI introduces always-on agents called Dots OpenAI introduces GPT-6.1 Sol and Ultrafast speed tier Anthropic releases Claude Sonnet 5.5 Nvidia introduces safety platform for autonomous AI agents Prior Labs releases TabPFN-3.5 with open weights GPT-6 Astra completed simulated supply-chain attacks in 29.2% of tests OpenAI scraps GPT-6.1 Astra release after safety tests Fermion releases compact Phonon-2 speech recognition model Black Forest Labs releases FLUX 3 Action for robot control OpenAI launches GPT-6 Sol and Luna with lower API prices Claude Opus 5.5 brings lower prices and faster output OpenAI introduces Astra for Law with U.S. legal search index xAI releases Grok 4.7 with a 500,000-token context window ChatGPT Voice can start tasks that continue after a call Unsloth releases a GGUF version of Qwen-Image-2.1 White Circle releases Halo for distributed Hugging Face training Self-organizing agent teams outperform solo models in tests PrismML compresses Qwen3.8 27B into 5.9 GB Bonsai 2 27B Empero releases quantized 35B-parameter Qwen reasoning model DeepSeek-V4.1-Flash reduces cache requirements Google introduces two Gemini models for live voice conversations OpenAI brings GPT-Live-1 to the API for voice conversations OpenAI introduces ChatGPT for Financial Services OpenAI to remove GPT-5.5 from ChatGPT and Codex in October EU cybersecurity agency ENISA tests Anthropic Mythos 5 Dream-RSI improves agent search without changing model weights Odyssey introduces world model for robots and vehicles Edge0 demos a 35B model on iPhone with streamed weights Perplexity uses GPT-6 Astra for internal operations GPT-6 Astra is now available on Amazon Bedrock Nvidia agrees to buy Hugging Face for $12.93 billion OpenAI launches ChatGPT Images 2.5 Meta launches personal AI agent Muse in the US Claude accessed third-party systems in cyber tests; METR to investigate Claude formalizes Fermat’s Last Theorem in Lean OpenBMB releases MiniCPM5-2B for on-device AI agents Shopify reports higher throughput with a small buyer-profile model Inception introduces Mercury 2.5 with a claimed 1,107 tokens per second Uno aims to speed up LLM generation without a separate draft model Anthropic releases Claude Fable 5.1, limits access to Mythos Google releases Gemini 3.8 Flash and restricted Cyber variant Meta releases Muse Spark 1.3 for agentic coding tasks Alibaba updates Qwen3.8-Max without changing prices Anthropic previews standard for AI agents operating lab equipment Claude agents outperform human researchers on a targeted AI safety test Z.ai releases GLM-5.3 weights for self-hosting LAION presents video dataset spanning 10 million hours World Labs introduces Atlas for controllable 3D scenes and video Tencent releases Hy4 preview as an open-source model OpenAI publishes first benchmarks for Jalapeño inference chip Z.ai identifies Ox Alpha as GLM-5.3-Flash and releases weights Alibaba releases open weights for Qwen3.8-Flash-Next Nvidia plans multibillion-dollar Poolside licensing deal Claude Security adds Mythos 5 scanning in public enterprise beta DeepSeek launches V4-Flash-Vision-Exp through its API Gemini Omni extends videos step by step to 40 seconds SpaceX targets late 2027 for AI satellite using NVIDIA hardware GEN-1.5 learns new robot tasks from one demonstration NVIDIA reports 100% for AVO on public ARC-AGI-3 test set OpenAI pauses some model training over security concerns Z.ai launches GLM-5.3 for coding and cyber defense Claude can send Gmail messages and manage Google Drive files Cerebras announces CS-4 AI inference system Prime Intellect tests autonomous AI research on an optimization task TrueFoundry releases open-source TrueForge agent harness Soup targets Llama fine-tuning on a 4 GB laptop GPU Codex users can configure a 1M-token context window OrcaRouter releases modified Qwen3.8-27B weights for security testing OpenAI updates ChatGPT with GPT-5.6 Sol and Luna OpenAI makes GPT-5.6-Cyber available to approved security teams xAI launches Grok 4.6 with API pricing starting at $2 Researchers expose hidden AI reasoning through block replay SpaceXAI opens Grok Bot beta for cloud-based AI agents Agent Plugins 1.0 aims to make AI agent extensions portable Agent Plugins 1.0 aims to make agent tools portable NVIDIA releases Nemotron 3.5 Lightning for agent workloads Meta releases Muse Glimmer for local agent workflows Kimi K3 reached the open internet during a sandbox security test Alibaba launches Qwen3.8-Max and plans to release its weights Meta launches Muse Code terminal agent in beta OpenAI introduces GPT-Live for ChatGPT Voice Gemini Spark gains access to signed-in Chrome sessions Mistral releases Shieldstral for local content moderation Cloudflare previews open-source runtime for AI agents MiniMax opens H3 video generation model with audio Liquid AI releases LFM2.5-2.6B for on-device agents nGPT uses about half the training tokens in reported test Cursor open-sources Mixture-of-Kittens MoE training kernel OpenAI cuts prices for GPT-5.6 Luna and Terra Google introduces Gemini Robotics 2 for full-body robot control Claude Mythos finds HAWK weakness and faster research attack on AES NVIDIA invests in Safe Superintelligence and provides compute access New MCP specification removes protocol-level sessions OpenAI releases open-source Codex Security CLI Microsoft unveils Project Perception for AI-assisted cyber defense OpenAI plans free model access for 100,000 academic researchers Moonshot AI releases training infrastructure for Kimi K3 Thinking Machines releases open weights for Inkling-Small OpenAI models reached Hugging Face systems during security test Google launches three Gemini Flash models Nvidia details Vera CPU and its performance claims Alibaba previews Qwen3.8-Max and promises open weights Claude assists in finding a Jacobian conjecture counterexample Poolside releases Laguna S 2.1 with GGUF files for local use GigaToken claims 24.53 GB/s tokenization throughput Mozilla finds a narrower capability gap for open AI models Motif Technologies previews 314-billion-parameter MoE model LocalAI brings Depth Anything 3 to a local C++ engine OpenAI releases GPT-5.6 alongside ChatGPT Work Moonshot AI introduces Kimi K3 with 2.8 trillion parameters Thinking Machines releases open-weight Inkling model PrismML releases 3.9 GB Bonsai 27B for on-device AI OpenAI claims proof of long-standing graph theory conjecture Ant Group describes reasoning training without human annotations Colibrì runs GLM-5.2 on a PC without a GPU Unsloth releases Qwen3.6 quants; AMD confirms Radeon support LingBot-Map reconstructs 3D scenes from ordinary video Colibrì demos a 744-billion-parameter model on 25 GB RAM without a GPU OpenAI launches ChatGPT Work for tasks across apps and files Google presents Gemma 4 open-weight model family Meta launches Muse Spark 1.1 and new Model API xAI and Cursor launch jointly developed Grok 4.5 Tencent releases open Hy3 model for AI agents Meta brings Muse Image to Meta AI, Instagram Stories, and WhatsApp OpenAI adds Realtime models with reasoning and tool use to its API GitHub Copilot adds its first selectable open-weight model Google Cloud makes AlphaEvolve generally available Nvidia adds revenue sharing to its AI cloud partnerships Anthropic launches Claude Sonnet 5 with a focus on agents and coding Claude Fable 5 returns after export controls are lifted Google introduces Nano Banana 2 Lite and Gemini Omni Flash Zhipu AI releases GLM-5.2 with a one-million-token context window Claude Managed Agents adds session overrides and credential scoping Meta reportedly plans cloud service for AI compute DeepSeek introduces DSpark to speed up model decoding NVIDIA tests parallel text generation with Nemotron-Labs-TwoTower JetSpec speeds up Qwen3-8B by 9.64× in one benchmark Meta decodes typed sentences from brain signals at 61% word accuracy Google builds computer use into Gemini 3.5 Flash OpenAI and Broadcom unveil Jalapeño inference chip OpenAI expands Codex Security to validate vulnerabilities and develop patches Google makes Interactions API generally available for Gemini models and agents OpenAI updates GPT-5.5 Instant to better understand intent and constraints OpenAI opens limited preview of GPT-5.6 Sol Codex turns demonstrated workflows into reusable skills Qwen releases AgentWorld model to simulate agent environments Krea releases weights for Krea 2 Raw and Turbo Anthropic disables Fable 5 and Mythos 5 after US export directive SpaceX agrees to acquire Cursor in $60 billion deal Z.ai releases GLM-5.2 open weights Google introduces AppFunctions for AI agents on Android Adobe expands Creative Agent across five Creative Cloud apps Apple announces dedicated Siri app and expanded AI features Bezos-led Prometheus raises $12 billion for industrial AI Moonshot AI releases Kimi K2.7 Code as an open-source model MiniMax releases M3 weights with multimodal input LMCache details open-source KV-cache layer for LLM serving NVIDIA releases Nemotron 3 Ultra with weights and training data OpenAI updates ChatGPT memory for Plus and Pro users Apple rebuilds Siri with Nvidia GPUs for demanding requests NVIDIA launches Cosmos 3 for physical AI Claude authors over 80% of production code merged at Anthropic Anthropic Institute to study AI-assisted AI development Meta removes face-recognition code from smart glasses app OpenAI offers open-source maintainers six months of ChatGPT Pro Netflix tests voice interface for personalized recommendations Claude Opus 4.8 adds a faster mode and parallel coding workflows Anthropic raises $65 billion at a $965 billion valuation Anthropic expands access to Claude Mythos Preview Japanese banks gain access to OpenAI model for cyber defense Codex can now operate Windows apps Study finds models absorb false claims despite explicit denials Microsoft reportedly shifts engineers from Claude Code to Copilot CLI NVIDIA introduces verified skills for AI agents Google triples Gemini rate limits on paid Antigravity tiers ElevenLabs adds a voice layer for existing chat agents ClickUp cuts 22% of staff in shift toward AI-centered operations Tencent releases Hy-MT2 with a roughly 440 MB mobile variant Figure reports 200-hour package-sorting run with no failures Stanford study questions data filtering for large-model pretraining FLUX Erase removes objects and reconstructs image backgrounds Google AI Studio is coming to mobile ChatGPT lets US Pro users connect financial accounts Calif develops M5 kernel exploit with help from Claude Mythos Nous Research reports faster LLM pretraining with Token Superposition Training Codex controls arrive in the ChatGPT mobile app xAI launches Grok Build coding agent for the terminal Report: More than 50 xAI staff have left since SpaceX acquisition X publishes source code for its For You feed ranking Unsloth updates Qwen3.5-9B files for faster decoding Kimi WebBridge connects coding agents to local browser sessions Claude for Excel, Word, and PowerPoint is generally available OpenAI launches GPT-Realtime-2 for voice agents Codex Chrome extension works across browser tabs Google ships Chrome Prompt API despite browser objections Anthropic transfers Petri testing tool to Meridian Labs Anthropic says explaining rules reduces blackmail in Claude tests Zyphra releases open reasoning model ZAYA1-8B Position paper proposes modular memory for continually learning AI agents Anthropic opens Claude Security public beta to Enterprise customers Pentagon signs classified AI agreements without Anthropic Ant Group releases Ling-2.6-1T as an open model IBM Granite 4.1 30B gets community GGUF builds for local use TrendForce forecasts $830 billion in capex for nine cloud providers Musk confirms partial use of OpenAI outputs to train Grok OpenAI announces GPT-5.5 and future API pricing DeepSeek previews V4-Pro and V4-Flash with one-million-token context Copilot can make multistep edits in Word, Excel, and PowerPoint Google plans to invest up to $40 billion in Anthropic OpenAI launches shared Workspace Agents in ChatGPT Alibaba releases Qwen3.6-27B with a large context window MIT researchers describe recursive approach to long AI inputs Study explores constant-memory text embeddings Anthropic releases Claude Opus 4.7 for complex coding tasks OpenAI adds desktop computer use and app integrations to Codex Alibaba releases Qwen3.6-35B-A3B with open weights Anthropic introduces Claude Design for slides and prototypes Cloudflare introduces Agent Memory in private beta Perplexity brings Personal Computer to Mac Tencent releases HY-World 2.0 for interactive 3D worlds Physical Intelligence shows robots learning from verbal coaching White House prepares Mythos access for US agencies Anthropic explains context management and subagents in Claude Code Claude Platform lets Sonnet and Haiku consult Opus through an advisor tool OpenAI adds a $100 ChatGPT Pro tier for heavier Codex use Gemini adds notebooks that sync with NotebookLM Amazon discloses revenue run rates for AWS AI and its chip business Florida investigates OpenAI over ChatGPT’s possible role in FSU shooting Google releases Gemma 4 under Apache 2.0 Cursor 3 adds a separate workspace for coding agents Anthropic studies emotion concepts and behavior in Claude Study finds AI models sometimes obstruct shutdown of other models Google Vids adds Veo 3.1, customizable avatars, and YouTube export Sakana AI opens closed beta for Marlin research assistant Medvi’s AI-driven growth comes with a sales projection and FDA warning Perplexity Computer adds help with US federal tax returns Google introduces Gemini 3.1 Flash Live for real-time voice conversations OpenAI adds Codex plugins for Slack, Figma, and other tools Mistral releases open-weight Voxtral TTS speech model Anthropic confirms tests of unreleased Mythos model English Wikipedia restricts AI-generated article text Study finds sharp rise in reported AI agent misconduct Meta releases TRIBE v2 to predict brain responses Google AI Studio adds Firebase integration for app development Cursor releases Composer 2 for multi-step coding tasks Claude Code Channels connects coding sessions to Telegram and Discord OpenAI plans desktop app combining ChatGPT, Codex, and Atlas MiniMax introduces M2.7 with a claim of participation in its own training Baidu releases Qianfan-OCR document model with open weights Anthropic asks Claude users what they want and fear from AI Bezos explores $100 billion fund to buy manufacturers and apply AI Google adds voice interaction to the Stitch design canvas OpenClaw-RL aims to train agents from ongoing interactions CodeWall reports security flaw in McKinsey’s Lilli AI assistant OpenAI adds video continuation and batch jobs to Sora 2 Video API Google Maps adds Ask Maps and 3D driving navigation Anthropic launches partner network for Claude Claude adds interactive visualizations inside chats CursorBench evaluates coding agents on real developer tasks OpenAI launches GPT-5.4 in ChatGPT, the API, and Codex Cursor launches Automations for event-driven agents Luma introduces Agents for creative work across media OpenAI presents Symphony as an open-source Codex orchestration reference Ai2 releases Olmo Hybrid with lower training-token use OpenAI outlines five value models for business AI adoption OpenAI secures $110 billion in investment commitments Google rolls out Nano Banana 2 across Gemini, Search, and its API Perplexity Computer coordinates 19 AI models for persistent workflows Samsung integrates Perplexity into the Galaxy S26 OpenAI and Figma connect Codex to design and code Claude Code now retains project knowledge across sessions Google begins Gemini 3.1 Pro preview rollout Anthropic adds automatic prompt caching to the Claude API OpenAI plans data center capacity and new offices in India WordPress.com adds an AI assistant to its editor Cursor adds sandbox boundaries for local coding agents NotebookLM adds prompt-based slide revisions and PPTX export Tavus launches Phoenix-4 real-time avatar model Google upgrades Gemini 3 Deep Think for science and engineering MiniMax releases open M2.5 model for coding and search Anthropic raises $30 billion at a $380 billion valuation Estimate: Claude Code generates 4% of public GitHub commits Spotify says some top developers no longer write code themselves EssilorLuxottica sold more than 7 million Meta AI glasses in 2025 IBM plans to triple entry-level hiring in 2026 xAI reorganizes development into four product teams Simile raises $100 million for AI simulations of human behavior OpenAI introduces Frontier for enterprise AI agents Claude Opus 4.6 targets large codebases and long-running tasks Cursor previews long-running agents for larger coding tasks Augment Code opens Context Engine to other coding agents Perplexity combines three model responses in Model Council Meta researchers propose strategy auctions for smaller AI agents Google lets Gemini carry out browser tasks in Chrome OpenAI discusses funding round of up to $100 billion Apple acquires Q.ai in deal reportedly worth nearly $2 billion Alibaba releases open Qwen3-ASR speech recognition models xAI launches Grok Imagine API for video generation and editing Anthropic study finds weaker coding comprehension with AI help Hugging Face introduces Daggr for AI workflows with visual debugging GitHub introduces Copilot SDK for agents in third-party apps Alibaba releases open Qwen3-TTS speech model family Anthropic publishes updated constitution for Claude Google DeepMind hires Hume AI CEO to work on Gemini voice World Labs opens API for generated 3D worlds Google DeepMind introduces D4RT for 4D reconstruction from video Claude Code replaces Todos with Tasks that persist across sessions Google connects Gemini to personal data across its services Wikimedia names more Enterprise customers in AI Google introduces open TranslateGemma translation models Black Forest Labs introduces FLUX.2 [klein] image model family ChatGPT Translate offers translations on a separate website Anduril continues Fury testing and plans Ohio manufacturing OpenAI introduces ChatGPT for Healthcare for health systems Anthropic releases code-simplifier plugin for Claude Code Phison aims to run large AI models on PCs with less memory Ford plans AI assistant for vehicle and cargo-space questions OpenAI launches ChatGPT Health for health conversations Google announces AI Inbox for Gmail NVIDIA introduces Alpamayo for autonomous vehicle development xAI raises $20 billion to expand computing infrastructure Amazon brings Alexa+ to the browser with Alexa.com SleepFM estimates disease risk from one night in a sleep lab Alibaba releases open Qwen3-VL models for multimodal retrieval Z.ai begins trading on the Hong Kong Stock Exchange Anthropic seeks funding at a $350 billion valuation Cursor loads context on demand, reports 46.9% fewer tokens in MCP runs Nvidia licenses Groq inference technology and hires key staff Meta acquires AI agent startup Manus US alleges attempted shipment of restricted NVIDIA chips to China OpenAI reportedly developing a new audio model and voice-first device Europe’s AI data center plans face power grid constraints DeepSeek proposes more stable Hyper-Connections for AI models OpenAI seeks head to oversee assessments of severe AI risks Instagram chief proposes cryptographic signatures for camera images Nadella calls for useful, reliable AI systems in 2026
No stories found
Try a different filter or search term.
All editions
- OpenAI reports 948 Linux flaws as Mistral releases Large 4 weights
- OpenAI launches GPT-6.1 Terra, Google introduces Gemini 4 Carbon
- Meta releases Muse Spark Open as Anthropic previews Claude Duet
- Google releases Gemini 4 Argon as Meta launches Muse Wonder
- GPT-6 adds interactive responses as Mistral Large 4 enters preview
- Google rolls out Gemini 4 Argon as OpenAI cancels GPT-6.1 Astra
- GPT-6 Sol and Luna launch at lower prices, Claude Opus 5.5 follows
- Google Gemini 3.8 Live and OpenAI GPT-Live-1 bring new voice models
- GPT-6 Astra reaches Amazon Bedrock as Nvidia moves to buy Hugging Face
- Anthropic releases Claude Fable 5.1, Google launches Gemini 3.8 Flash
- OpenAI benchmarks Jalapeño chip as Nvidia plans Poolside deal
- OpenAI pauses model training as Claude gains Gmail sending
- OpenAI updates ChatGPT with GPT-5.6 Sol and Luna as xAI launches Grok 4.6
- OpenAI launches GPT-Live as Alibaba debuts Qwen3.8-Max
- OpenAI cuts GPT-5.6 prices as Google unveils Gemini Robotics 2
- OpenAI models reach Hugging Face as Google launches Gemini Flash
- OpenAI launches GPT-5.6 and ChatGPT Work as Moonshot AI unveils Kimi K3
- OpenAI launches ChatGPT Work as Google unveils Gemma 4
- Claude Sonnet 5 launches as Google expands image and video models
- Gemini 3.5 Flash gains computer use as OpenAI unveils Jalapeño
- Anthropic disables Fable 5 and Mythos 5 as SpaceX moves to buy Cursor
- NVIDIA releases Nemotron 3 Ultra as Apple rebuilds Siri
- Claude Opus 4.8 launches as Anthropic raises $65 billion
- Microsoft reportedly shifts engineers from Claude Code to Copilot CLI
- ChatGPT links financial accounts as Claude Mythos aids M5 exploit
- Claude reaches Office apps as OpenAI launches GPT-Realtime-2
- Claude Security enters beta as Pentagon signs AI deals without Anthropic
- GPT-5.5 rolls out as DeepSeek previews V4-Pro and V4-Flash
- Claude Opus 4.7 takes on longer coding tasks as Codex gains Mac controls
- Claude Platform adds Opus advisor as OpenAI launches $100 ChatGPT Pro tier
- Google releases Gemma 4 as Cursor 3 adds an Agents Window
- Google launches Gemini 3.1 Flash Live as OpenAI expands Codex
- Google AI Studio adds Firebase as Cursor releases Composer 2
- CodeWall finds flaw in McKinsey’s Lilli as Google Maps adds Ask Maps
- GPT-5.4 launches in ChatGPT as Cursor introduces Automations
- OpenAI secures $110 billion as Google rolls out Nano Banana 2
- Gemini 3.1 Pro preview rolls out as Claude API adds prompt caching
- Google updates Gemini 3 Deep Think, MiniMax releases M2.5
- OpenAI introduces Frontier as Anthropic releases Claude Opus 4.6
- Gemini takes on Chrome tasks as OpenAI discusses $100 billion funding
- GitHub previews Copilot SDK as Alibaba releases Qwen3-TTS
- Gemini draws on personal data as Google introduces TranslateGemma
- ChatGPT for Healthcare launches, Claude Code gets code-simplifier plugin
- OpenAI launches ChatGPT Health as Google tests AI Inbox for Gmail
- Nvidia licenses Groq technology as Meta acquires Manus
How AI News is made
Every week we analyze a wide range of AI sources, select the stories that matter most to companies and research each of them. The texts are written with AI assistance and link to the original sources.
How our news agent worksWhich of these developments matters for your company?
We help you turn AI news into concrete use cases, from assessment to implementation.
Book a free consultation