AI LLM News
AI LLM NewsAlibaba’s Qwen3.8-27B Tests the Open Multimodal Hype Cycle
A 27B rumour is not the Qwen3-VL GroundSlate already runs.
14 Aug 2026
Gemini 3.7 Flash Is Google's New Speed Play for AI Automation
Gemini 3.7 Flash is Google's cheaper, faster automation tier.
13 Aug 2026
AI LLM NewsGrok 4.6 Pushes AI Agents Toward Real Work
Grok 4.6 moves xAI into agent workflows you can actually run.
12 Aug 2026
Archive
- Rent Opus 5 for ugly tickets. Keep Kokoro31 Jul 2026
- Moonshot AI's Kimi K3 Pushes Open-Weight AI Into Frontier Territory29 Jul 2026
- Claude Opus 5 Turns Long Context Into Workflow Muscle27 Jul 2026
- Claude Fable 5 Is Back, With a Credit Meter Attached6 Jul 2026
- LongCat 2.0 Pushes Open-Source AI Toward Million-Token Workflows5 Jul 2026
- Claude Sonnet 5 Moves AI From Chat to Workflow2 Jul 2026
- OpenAI's GPT-5.6 Preview Turns Model Choice Into Workflow Strategy28 Jun 2026
- Claude Mythos 5 Returns Under Lock and Key27 Jun 2026
- Google Gemini 3.5 Flash Gets Native Computer Use25 Jun 2026
- xAI Grok 4.3 Pushes Into Long-Context Ops With 1M Tokens and API Access7 May 2026
- OpenAI’s GPT-5.5 Instant Is Here, and the Real Upgrade Is Workflow Speed With Fewer Weird Moments5 May 2026
- Google Pushes Gemini 3.1 Ultra to 2M Tokens, and That Changes the Workflow Math4 May 2026
- Mistral Medium 3.5 Wants to Be the Open Agent Model Adults Can Actually Deploy3 May 2026
- NVIDIA Nemotron 3 Nano Omni Makes Multimodal AI More Operational1 May 2026
- DeepSeek V4 Brings 1M-Context Open Weights Into the Automation Race25 Apr 2026
- GPT-5.5 Pushes OpenAI Deeper Into Real Agent Work24 Apr 2026
- Moonshot AI’s Kimi K2.6 Pushes Open Models Closer to Real Agent Work22 Apr 2026
- ChatGPT Images 2.0 Pushes AI Visuals Closer to Real Workflow Territory21 Apr 2026
- Anthropic Introduces Claude Opus 4.7: Reliability Becomes the Product18 Apr 2026
- Anthropic has introduced Claude Design as a research preview17 Apr 2026
- OpenAI has released gpt-oss-20b and gpt-oss-120b14 Apr 2026
- Anthropic’s Claude Mythos Is Real. The Open API Still Isn’t.11 Apr 2026
- Meta’s Muse Spark Wants to Be More Than a Chatbot8 Apr 2026
- Qwen3.6-Plus Wants to Be the Agent Brain, Not Just Another Chatbot6 Apr 2026
- GLM-5V-Turbo Turns Screens Into Code, but the API Story Is What Makes It Matter4 Apr 2026
- Google DeepMind’s Gemma 4 Is Open for Business3 Apr 2026
- Alibaba’s Qwen3.6-Plus Pushes Multimodal AI Closer to Real Agent Work2 Apr 2026
- François Chollet’s ARC-AGI-3 Is Here, and It’s a Brutal Reality Check for “Agentic” AI31 Mar 2026
- Platform Lock-In Is Coming for Your AI Stack30 Mar 2026
- Anthropic’s Claude Mythos Leak Is Real. The API Story Isn’t.27 Mar 2026
- Gemini 3.1 Capabilities18 Mar 2026
- Kimi 2.5 Agent Swarm18 Mar 2026
- OpenAI’s GPT-5.4 Mini and Nano: Small Models, Big Automation Energy18 Mar 2026
- Anthropic Dispatch Turns Claude Into Your Always-On Creative Coworker17 Mar 2026
- NVIDIA’s Nemotron-3 Super (120B) Lands as a “Callable” Model for Enterprise Agents16 Mar 2026
- Grok-4.20 Beta Debuts as xAI’s New Reasoning Powerhouse13 Mar 2026
- Grok 4.20 Beta Makes Multi-Agent AI the Default (and That’s a Workflow Shift)9 Mar 2026
- Sarvam AI Releases Open-Weights LLMs Built for India’s Language Reality7 Mar 2026
- OpenAI Launches GPT-5.4 With Configurable Reasoning for Automation5 Mar 2026
- Google’s Gemini 3.1 Flash-Lite Is a Speed-and-Cost Upgrade Built for Real Automation3 Mar 2026
- DeepSeek V4 Teases a 1M Token, Multimodal All in One Model. Here is What is Actually Operational1 Mar 2026
- Meta’s Llama 3.3 (70B): The Less Sexy, More Useful Model Release28 Feb 2026
- HyperNova 60B Isn’t Just “Open Weight.” It’s On-Prem Agent Fuel.26 Feb 2026
- Inception Labs’ Mercury 2 Makes Text AI Feel Live (and That’s the Point)24 Feb 2026
- Google Gemini 3.1 Pro Lands as a Reasoning Upgrade You Can Actually Automate23 Feb 2026
- Google’s Gemini 3.1 Pro Is a Reasoning Upgrade and a Workflow Upgrade Too21 Feb 2026
- Alibaba’s Qwen3.5-397B-A17B Drops: Open-Weights, Multimodal, and Agent-Ready (Finally the Combo Ops Teams Want)19 Feb 2026
- Sarvam AI Drops India-Focused LLMs (105B + 30B) and Yes, You Can Actually Use Them19 Feb 2026
- Claude Sonnet 4.6: The 1M Token Upgrade That Turns “Chat” Into Ops17 Feb 2026
- OpenAI’s GPT-5.3-Codex-Spark Is a Speed Tier for Coding and a Hardware Pivot That Matters15 Feb 2026
- Google DeepMind’s Gemini 3 “Deep Think” Upgrade Is a Reasoning Flex and an Automation Tease14 Feb 2026
- Zhipu AI’s GLM-5 Is Open-Weights and Built for Agentic Work13 Feb 2026
- Claude Opus 4.6 Makes “One Session” Feel Like a Workspace11 Feb 2026
- OpenAI’s GPT-5.3-Codex Makes Agentic Coding Faster, Cheaper, and (Finally) Interruptible6 Feb 2026
- Anthropic Claude Opus 4.6 Pushes Long-Context + “Agent Teams” Into Real Workflow Territory5 Feb 2026
- Qwen3-Coder-Next-3B: Alibaba’s Lean MoE Model Aims at Real Coding Automation3 Feb 2026
- Moonshot AI’s Kimi K2.5 Drops “Agent Swarm” Parallelism and It’s Built for Automation29 Jan 2026
- Zhipu AI’s GLM-4.7-Flash Is the “Run It Yourself” Speed Drop28 Jan 2026
- TranslateGemma 27B Makes Translation Deployable21 Jan 2026
- ChatGPT Ads Test: New Intent Surface for Marketers16 Jan 2026
- Tencent HY-MT1.5 Brings Offline Translation to Workflows14 Jan 2026
- MiniMax-M2.1: Open Weights Built for Real Agents12 Jan 2026
- Liquid AI LFM2.5 Makes On Device Speech Agents Real6 Jan 2026
- Gemini in Google Ads: Campaigns Start via Chat20 Dec 2025
- Salesforce Agentforce Sales Lands Inside ChatGPT19 Dec 2025
- ChatGPT Apps Directory Makes Chat a Work OS18 Dec 2025
- Nemotron-3 Makes Open Agentic AI Production-Ready16 Dec 2025
- LLM Powerhouses: GPT-5.2, Gemini, DeepSeek Transform Workflows12 Dec 2025
- Gemini 3 vs ChatGPT: Future-Proof Your AI Stack25 Nov 2025
- Google Gemini 3 Arrives: What’s Real Now?18 Nov 2025
- Qwen, Grok, GPT‑5.1: AI Assistants Get Real18 Nov 2025
- ChatGPT 5.1 Instant vs Thinking Explained13 Nov 2025
- Emu3.5 Lands: Multimodal Ambition Meets Reality1 Nov 2025
- Claude Haiku 4.5: Fast, Affordable Automation31 Oct 2025
- NVIDIA Nemotron Nano 2: Practical Automation at Speed29 Oct 2025
- Alibaba Cloud Launches Qwen3-Max: Trillion-Parameter LLM19 Oct 2025
- Alibaba’s Qwen3‑Max: API-Ready Trillion Parameter Model19 Oct 2025
- GLM‑4.6: Open Weights, 200K Context, Real Automation13 Oct 2025
- Claude Sonnet 4.5: Real Work, Real Automation1 Oct 2025
- Cloudflare flips JSON Mode for Workers AI: structured outputs at the edge20 Sept 2025
- Ollama Ships ARM Builds and Hardware Acceleration: Local AI Gets a Major Speed Boost16 Sept 2025
- Grok Code Fast 1: xAI’s Speed Play for Agentic Coding and What It Unlocks for Creative Ops2 Sept 2025