Latest news from the AI world
On April 20, 2026, a busy day: Anthropic and Amazon sign a major 5 GW and $5B deal, GitHub pauses new Copilot Pro/Pro+ signups, Kimi K2.6 sets new open-source coding records, and Qwen previews its next flagship model.
April 18, 2026: xAI launches audio STT and TTS APIs at the best market prices, Claude for Word in beta for Pro/Max/Team/Enterprise, Midjourney V8.1 with native 2K rendering, Luma opens Innovative Dreams with AWS, and Kimi publishes its PraaS architecture.
April 17, 2026: Claude Design in research preview, Claude Code v2.1.113 in native binary, Codex CLI v0.121 with plugin marketplace, Copilot CLI Auto Model GA, NotebookLM Free in Gemini App, NVIDIA NemoClaw and Runway Seedance 2.0 iOS.
April 16, 2026: Claude Opus 4.7 GA with SWE-bench Pro 64.3%, Codex extends computer use to macOS, GPT-Rosalind for the life sciences, Gemini integrates Nano Banana 2 into Google Photos, Perplexity Personal Computer, and the gh skill command.
Anthropic redesigns Claude Code on desktop with parallel sessions, OpenAI launches Agents SDK v0.14.0 with native sandbox, Google releases Gemini 3.1 Flash TTS and its macOS app on April 15, 2026.
April 14, 2026: Anthropic launches routines in Claude Code (planning, API, webhooks), Google DeepMind releases Gemini Robotics-ER 1.6 with industrial instrument reading, Z.ai opens GLM-5.1 under MIT license, and GitHub Copilot gains three new features.
April 13, 2026: GitHub launches remote control for Copilot CLI sessions from web and mobile, MiniMax releases M2.7 open-source with day-0 vLLM support, the Qwen3.5-Omni API becomes available internationally, and Gemini 3.1 Flash Live takes the lead in the τ-Voice voice leaderboard.
April 11, 2026: Claude Code 2.1.101 with system CA certificates and dynamic /loop, MiniMax launches Music 2.6 and MMX-CLI, GitHub suspends Copilot Pro trials following abuse, Qwen Code v0.14 adds Channels and Cron Jobs.
April 10, 2026: OpenAI details the Axios supply chain incident, GitHub Copilot CLI supports BYOK and offline local models, and ElevenLabs rolls out its voice AI on-premise and on-device.
Week of April 5 to 9, 2026: Anthropic launches Project Glasswing with Claude Mythos Preview (SWE-bench 93.9%), Meta unveils Muse Spark, Gemini CLI v0.37.0, Stability AI Brand Studio, and more.
April 3-4, 2026: Anthropic removes third-party tool access from Claude Pro and Max subscriptions with compensatory measures, publishes Claude Code v2.1.92 with a Bedrock wizard and per-model cost details. Google launches Veo 3.1 Lite in the API, Lyria 3 in the Gemini app, and expands video generation in Workspace.
On April 3, 2026, Anthropic opens Microsoft 365 connectors to all Claude plans, GitHub strengthens the Copilot cloud agent security with signed commits and organization-level firewall, and Gemini CLI moves to version 0.36.0 with native sandboxing.
April 2, 2026: Google releases Gemma 4 under Apache 2.0 with native vision and audio, Qwen3.6-Plus tops Terminal-Bench with a 1M-token context window, and Anthropic publishes foundational research on functional emotions in language models.
April 1, 2026: Claude Code v2.1.89 with NO_FLICKER and Computer Use, Copilot CLI /fleet for parallel agents, Perplexity Computer integrated in Slack, Codex CLI 0.118.0, GLM-5V-Turbo from Z.ai, GrandCode beats Codeforces, Runway Builders, and two April Fools' jokes.
March 31, 2026: OpenAI closes a record financing round at a $852 billion valuation, Qwen launches its native omnimodal model, and Boris Cherny reveals little-known features of Claude Code.
Week of Mar 26–28, 2026: Luma launches Uni-1 (unified visual reasoning), Perplexity powers Samsung Browsing Assist (1B+ devices), Claude Code v2.1.86 brings ~15 major fixes, GitHub Copilot CLI adds agent-driven unit test generation.
March 27, 2026: Amsterdam court fines xAI for non-consensual nudification (€100,000/day), Gemini 3.1 Flash Live real-time audio in 90+ languages, Codex plugins for Slack/Figma/Notion, Claude Code v2.1.85 and auto mode, Cohere Transcribe #1 ASR, Suno v5.5, Voxtral TTS.
March 25–26, 2026: Claude integrates work tools on mobile (Figma, Canva, Amplitude), Claude Code starts 2.8x faster thanks to 60 days of Bun optimizations, GitHub Copilot changes its interaction data policy with an opt-out to enable before April 24, and Lyria 3 Pro generates music tracks up to 3 minutes.
March 24, 2026: Claude Code launches auto mode with an action classifier, ChatGPT adds visual shopping with the Agentic Commerce Protocol, and xAI opens Grok Imagine to multi-image video generation on API.
Midjourney V8 Alpha 5x faster in native 2K, OpenAI buys Astral (uv, Ruff) for Codex, NVIDIA OpenShell open source runtime for autonomous agents. Week of March 18-23, 2026.
17 March 2026: OpenAI launches GPT-5.4 mini and nano with benchmarks close to the full model, Mistral and Perplexity join the NVIDIA Nemotron Coalition at GTC 2026, and Perplexity opens Comet to enterprise teams with MDM governance.
March 12–14, 2026: Anthropic opens the 1M token context window generally available for Opus 4.6 and Sonnet 4.6, Genspark unveils AI Workspace 3.0 with its agent Claw and $385M raised, Perplexity Computer arrives on smartphones.
March 12, 2026: Claude generates interactive charts and diagrams inline in conversation across all plans, OpenAI opens the Video API with Sora 2 to developers, and Google Maps receives its biggest overhaul in ten years thanks to Gemini.
March 11, 2026: Anthropic creates an interdisciplinary institute led by Jack Clark, Perplexity launches its 'Computer' vision with four APIs, OpenAI equips the Responses API with a full computing environment for agents, and Meta publishes four generations of MTIA chips.
March 10, 2026: Claude Code launches Code Review in research preview, OpenAI acquires Promptfoo for enterprise AI security, Runway opens its GWM-1 Characters video agents API.
March 8, 2026: Claude Code v2.1.71 introduces /loop to schedule recurring tasks, GitHub Copilot VS Code v1.110 brings agent plugins and browser agent tools, Pika launches AI Selves for video digital twins, and Luma unveils its creative agents.
6 March 2026: Anthropic and Mozilla reveal 22 vulnerabilities in Firefox thanks to Opus 4.6, OpenAI launches Codex Security with 14 CVEs detected in major open source projects, and Kling 3.0 deploys Motion Control worldwide.
Week of March 3–5, 2026: OpenAI's GPT-5.4 launches native computer use and reaches 75% on OSWorld, NotebookLM generates immersive videos with Gemini as director, Codex arrives on Windows, and GitHub enables Copilot Memory by default.
March 3, 2026: Claude Code launches Voice Mode (push-to-talk, free), GPT-5.3 Instant replaces GPT-5.2 in ChatGPT with hallucinations down 26.8%, Gemini 3.1 Flash-Lite in preview. Teaser GPT-5.4, FLUX.2 [pro] 2x faster, ElevenLabs at MWC.
March 2, 2026: Claude Code v2.1.63 delivers /simplify, HTTP hooks and 8 memory fixes, OpenAI signs a classified AI agreement with the Pentagon, Anthropic refuses mass surveillance and threatens to sue. Also: Claude Memory on the free plan, Qwen 3.5 Small Series, Gemini CLI v0.31.0, and deprecations of Gemini 3 Pro and GPT-5.1 in Copilot.
February 27, 2026: OpenAI announces a $110 billion round with SoftBank, NVIDIA and Amazon, an exclusive AWS cloud partnership, Gemini Drop with Lyria 3 and Nano Banana 2, and GitHub launches Agentic Workflows in Markdown.
26 February 2026: Dario Amodei publishes a statement on Anthropic's limits vis‑à‑vis the U.S. Department of War, Perplexity obtains OS-level access on the Samsung Galaxy S26, Claude and Codex go GA for Copilot Business and Pro, NVIDIA unveils LillyPod with 1,016 Blackwell Ultra GPUs.
February 25, 2026: Anthropic buys Vercept to accelerate computer use (Claude Sonnet 4.6 at 72.5% on OSWorld), Perplexity launches Computer — an autonomous multi-model orchestrator, GitHub Copilot CLI goes generally available with MCP and agents support.
February 24, 2026: Anthropic publishes the third version of its Responsible Scaling Policy with a public Frontier Safety Roadmap and quarterly Risk Reports; GPT-5.3-Codex moves to general availability via the Responses API; Meta signs a multi-year AMD deal for ~6 GW of GPU capacity; Qwen 3.5 Medium launches 4 MoE models including a 35B that outperforms the 235B giant; Claude Code v2.1.51 introduces Remote Control from mobile.
February 23, 2026: Anthropic reveals that DeepSeek, Moonshot AI and MiniMax extracted 16 million exchanges via 24,000 fraudulent accounts to distill Claude; OpenAI drops SWE-bench Verified (59.4% defective tests) and recommends SWE-bench Pro; gpt-realtime-1.5, WebSockets in the Responses API, Veo 3.1 templates and Pika AI Selves.
Claude Code v2.1.49-2.1.50 extends multi-worktree agents and fixes 6 memory leaks, GitHub Copilot introduces cross-session persistent memory, and ElevenLabs cuts TTS latency by 40% on February 22, 2026.
Anthropic launches Claude Code Security in preview (500+ vulnerabilities detected in open source), Runway becomes a multi-model hub, and GitHub Copilot extends its model picker to Business/Enterprise on February 20, 2026.
Google launches Gemini 3.1 Pro with an ARC-AGI-2 score of 77.1%, OpenAI expands to India with Tata Group and Stargate infrastructure, Anthropic opens Claude in PowerPoint to Pro plans, and Mistral publishes its Voxtral Realtime technical report.
February 16-18, 2026: Anthropic launches Claude Sonnet 4.6 with improved computer use and 1M token window. Qwen releases Qwen3.5-397B in open-weight Apache 2.0. Google integrates Lyria 3 into Gemini. OpenAI unveils EVMbench, Cohere launches Tiny Aya, Runway opens its Gen-4.5 API.
February 14-16, 2026: Anthropic opens Bengaluru office with major partnerships in India. Kimi unveils Claw, a native OpenClaw app with 5,000+ skills. GitHub Copilot integrates MCP. OpenAI publishes Codex recap and launches First Proof challenge.
February 1-14, 2026: MiniMax releases M2.5, open-source frontier model with 80.2% on SWE-Bench. Kling launches its 3.0 model for video. Perplexity deploys Model Council (multi-model), Deep Research on Opus 4.6, and DRACO benchmark. Mistral announces its biggest hackathon, Cohere signs Magnus Carlsen.
February 13, 2026: GitHub launches Agentic Workflows in technical preview, GPT-5.2 derives a result in particle physics, OpenAI introduces Lockdown Mode against prompt injections, and Anthropic and Google invest in AI education.
Anthropic announces a $30 billion Series G ($380B valuation), OpenAI launches real-time GPT-5.3-Codex-Spark on Cerebras, Google unveils Gemini 3 Deep Think with record scores in sciences, and xAI restructures teams with Macrohard.
Z.ai launches GLM-5 (744B params, MIT License), Anthropic releases its first ASL-4 sabotage risk report, OpenAI adds server-side compaction and network containers to its API, Kimi unveils Agent Swarm, and Runway raises $315M.
Anthropic launches 2.5x faster mode for Opus 4.6, OpenAI tests ads in ChatGPT, ElevenLabs closes $11B Series D, GPT-5.3-Codex arrives in Copilot, and Mistral releases Voxtral Transcribe 2.
GitHub opens Agent HQ to Claude and Codex in public preview, Anthropic explains why Claude will remain ad-free, Apple integrates Claude Agent SDK into Xcode, Mistral launches Voxtral Transcribe 2, and ElevenLabs raises $500M in Series D.
Anthropic launches Claude Opus 4.6 with 1M context and agent teams, OpenAI releases GPT-5.3-Codex and the Frontier platform, Google updates Gemini 3 for the Super Bowl, and GitHub introduces pinned comments.
SpaceX acquires xAI to deploy AI in space, OpenAI releases Codex App on macOS, Qwen releases an 80B model for coding agents, Anthropic partners with Allen Institute and HHMI for scientific research.
Google launches Project Genie to create interactive worlds, Gemini arrives in Chrome with Auto Browse, Veo 3.1 adds image-to-video, and Personal Intelligence connects your Google apps.
Claude plans the first AI drive on Mars, Anthropic enriches Cowork with 11 plugins and publishes a study on loss of autonomy, OpenAI accelerates ChatGPT apps, xAI launches a cutting-edge video API.
AI News January 26-28, 2026: Gemini comes to Chrome with auto browse, OpenAI launches Prism for scientists, Claude integrates interactive tools, Anthropic partners with UK government, Google AI Plus available in the US.
AI News Jan 21-23, 2026: Claude comes to Excel, Tasks system in Claude Code v2.1.19, Codex agent loop architecture, PostgreSQL for 800M ChatGPT users, Qwen3-TTS open-source, Runway Gen-4.5 Image to Video.
Anthropic publishes Claude's new constitution and appoints new member to Long-Term Benefit Trust, OpenAI launches Education for Countries and age prediction, Google rolls out free SAT practice and Classroom updates, Manus launches App Publishing
OpenAI launches ChatGPT Go at $8/mo in 170+ countries with ads, Anthropic opens India office with Irina Ghose, Google unveils MedGemma for medical, and ElevenLabs powers MasterClass On Call.
Anthropic launches Labs with Mike Krieger, OpenAI signs with Cerebras and Merge Labs, Google and Apple announce Gemini for Siri, ElevenLabs arrives at Deutsche Telekom, and Qwen releases its first safety model.
Google launches Personal Intelligence in Gemini for personalized answers. OpenAI makes GPT-5.2-Codex available via API. NotebookLM rolls out Data Tables to everyone.
xAI completes record $20B Series E funding. OpenAI acquires Torch to strengthen ChatGPT Health. Qwen launches multimodal models.
Anthropic launches Cowork, Claude Code for non-developers. New connectors and Agent Skills for the healthcare sector.
Anthropic publishes Constitutional Classifiers++ after 1700h of red-teaming. OpenAI launches enterprise offering for hospitals. ElevenLabs unveils Scribe v2 for transcription.
OpenAI launches ChatGPT Health. Z.ai goes public in Hong Kong. ElevenLabs invests $33M in NYC. Claude Code v2.1.0 with Skills.
NVIDIA unveils its new architectures at CES 2026: Vera CPU for agentic AI, Rubin with 6 chips, Alpamayo for autonomous vehicles, and major partnerships.
Early 2026: OpenAI launches Grove for founders, Your Year with ChatGPT rolled out, 2025 developer recap, and Qwen-Image-2512 available.
End of 2025: xAI launches Grok Business and Collections API, Manus joins Meta AI, WAN 2.6 improves video consistency, Runway and Adobe partner up.
Z.ai (Zhipu AI) launches GLM-4.7, new open-source leader surpassing GPT-5.1 on Vending-Bench2, and prepares for its IPO on January 8, 2026.
Weekly recap: Claude 2x usage promo, Qwen3-TTS with voice cloning, Mistral Vibe CLI Skills, ChatGPT Atlas security, and Gemini photo editing.
Qwen-Image-Edit-2511, Grok Collections API with integrated RAG, xAI-DOW partnership, NotebookLM Data Tables and Gemini CLI v0.22-23.
![FLUX.2 [max]: Black Forest Labs launches its most powerful model](/_astro/flux-2-max.DEpn9xsV_Sv3of.webp)
Black Forest Labs unveils FLUX.2 [max] with web contextual generation and support for 10 reference images. World #2 on ArtificialAnlys.
Generative AI news in December 2025: Luma Ray3 Modify, Pika API, Midjourney UI, ElevenLabs Music, Suno Personas and more.
Runway launches Gen-4.5 and details its Universal World Simulator vision. A video model that understands anatomy, physics, and motion.
December 2025 marks a major turning point for GitHub Copilot with the arrival of GPT-5.2, Claude Opus 4.5, and Gemini 3, plus Agent Skills and Copilot Memory.
Week of December 15-21, 2025: GPT-5.2-Codex for agentic coding, official Skills support in Codex CLI, GPT Image 1.5 in API, ChatGPT Apps directory, new audio models.
December 2025: Alibaba launches Qwen-Image-Layered, an open-source image decomposition model into RGBA layers with Photoshop-level quality.
Week of December 15-21, 2025: Claude in Chrome available to all, Claude Code integration, Bloom for AI evaluations, Project Vend Phase 2, DOE Genesis Mission partnership.
Recap week 3 December 2025: Gemini 3 Flash launched December 17, available for free in Gemini App, AI Mode, and Gemini CLI v0.21.0.
December 2025 recap: Devstral 2 (72% SWE-bench), Mistral Vibe CLI, Mistral 3 (Large 3, Ministral), Mistral OCR 3 for document processing.
xAI launches the OpenAI Realtime compatible Grok Voice Agent API, and announces a historic partnership with El Salvador to educate 1 million children with Grok.
All the new Claude Code features for December 2025: Slack integration, async subagents, Agent SDK updates, Android preview, plugins marketplace and more.
Week of December 8-14, 2025: GPT-5.2 with GDPval 70.9% and SWE-Bench 80%, Codex for Linear to automate issues, State of Enterprise AI report.
Week of December 8-14, 2025: Anthropic donates MCP to the Linux Foundation, creation of the Agentic AI Foundation with OpenAI and Block, Accenture partnership.
Week of December 1-7, 2025: Strategic acquisition of Bun, Claude Code hits $1 billion, $200M Snowflake partnership, and launch of Claude for Nonprofits.
GitHub Copilot 2025 Retrospective: The evolution towards a multi-model assistant with agents, MCP Server, improved CLI, and new intelligent editing capabilities.
2025 Retrospective: From Qwen2.5-Max to Qwen3-Coder, Alibaba Cloud establishes itself as a leader in open-source AI with models rivaling GPT-4 and Claude.
2025 Retrospective: xAI goes from Grok 3 Beta to Grok 4.1, conquers governments, and expands internationally.
Complete analysis of Claude Opus 4.5, Anthropic's new flagship model. Benchmarks, token efficiency, pricing, and enterprise feedback.
Retrospective September-November 2025: launch of Gemini CLI Extensions, Gemini 3 Pro, Nano Banana Pro, Deep Think mode, and 20 CLI versions.
Recap of Mistral AI announcements from September to November 2025: Le Chat with MCP connectors, Memories, AI Studio, 1.7B€ funding and European expansion.
Recap of Anthropic announcements from September to November 2025: Claude Sonnet 4.5, Haiku 4.5, Microsoft partnerships, Claude Desktop and more.