News
Headlines from publishers around the world, checked every few hours. Each link goes to the original story.
- OVHcloud unveils Olala mini, its first in-house LLMIT Brief UK, Releases, 2026-11-18
- A 9B model beats a 1T model on FinQA, trained with Crusoe Managed RLCrusoe, China, 2026-11-18
- The internal metaphysics of an LLMEurekAlert!, Releases, 2026-11-18
- Mapping Language Model Internal Representations to Physical Laws and Scientific ReasoningBioengineer.org, Releases, 2026-11-18
- Can China keep rogue AI in check? Meet the start-up testing the top modelsSouth China Morning Post, Research, 2026-11-18
- Microsoft aims to challenge OpenAI, Anthropic with cheaper, specialized AI modelsAxios, Frontier, 2026-11-17
- Microsoft rolls out new first-party models and updates to agentic AI continuous improvementSiliconANGLE, Open models, 2026-11-17
- Microsoft unveils AI models for coding, cybersecurity as it takes on Anthropic, OpenAIReuters, Frontier, 2026-11-17
- Google is making it easier to opt out of Gemini AI model training – here’s howZDNET, Frontier, 2026-11-17
- OpenAI Unveils New Astra AI Model After Scrapping Prior ReleaseBloomberg.com, Frontier, 2026-11-17
- AI safety legislation has a conflict-of-interest problem — Big US labs are warning of risk just as cheaper Chinese open-weight models erode their commercial edgeTom's Hardware, Research, 2026-11-17
- Anthropic, Other AI Models Regularly Identify Civilians as Military Targets in TestsBloomberg.com, Frontier, 2026-11-17
- OVHcloud launches European-trained open source AI modelIT Pro, Open models, 2026-11-17
- Reading and steering representations of materials-science mechanisms in an open-weight language modelEurekAlert!, Releases, 2026-11-17
- Moonshot, Tencent and Zhipu trim hacking skills from new Chinese AI modelsStartup Fortune, China, 2026-11-16
- China’s Top AI Firms Limit Cyber Capabilities in Latest ModelsBloomberg.com, China, 2026-11-16
- Alibaba's new Qwen AI model poses challenge for AnthropicYahoo Finance, China, 2026-11-16
- Alibaba's new Qwen AI model poses challenge for AnthropicYahoo Finance, China, 2026-11-16
- Alibaba Takes on Anthropic With New Qwen AI ModelPYMNTS.com, China, 2026-11-16
- Moonshot AI pivots Kimi models from coding to science as benchmark gains flattenThe Decoder, China, 2026-11-16
- California sues OpenAI over Hugging Face, other AI model loss of controlCNBC, Frontier, 2026-11-16
- OpenAI Brings Four Outside Research Groups Into AI Safety TestingBenzinga, Research, 2026-11-16
- Alibaba's Qwen4-Max takes on Opus 5.5, GPT-6, and Kimi K3.1 Pro at much lower pricesThe Decoder, China, 2026-11-16
- Jailbroken PS4 runs 9-billion-parameter AI model at up to 43 tokens per second — open-source Garlic LLM bypasses drivers to talk directly to the 2013-era GPUTom's Hardware, Open models, 2026-11-16
- Alibaba challenges OpenAI and Anthropic with Qwen4-MaxTechzine Global, China, 2026-11-16
- Alibaba Unveils New Flagship AI Model in Riposte to US RivalsYahoo Finance, China, 2026-11-16
- China's Zhejiang flags IP risks in commercial use of open-source AI modelsMLex, Open models, 2026-11-16
- China’s Moonshot AI puts Kimi models into race for scientific discoveriesSouth China Morning Post, China, 2026-11-16
- China’s Moonshot AI puts Kimi models into race for science discoveriesSouth China Morning Post, China, 2026-11-16
- Alibaba’s Giant Qwen4-Max Model Trails Claude, GPT and Moonshot AIwww.trendingtopics.eu, China, 2026-11-16
- Tether is pushing the 13-billion parameter BitNet b1.58 LLM to the edge.TechCrunch, Releases, 2026-11-16
- Alibaba launches Qwen4-Max model with strict cybersecurity guardrailsSouth China Morning Post, China, 2026-11-16
- Alibaba launches Qwen4-Max model with strict cybersecurity guardrailsSouth China Morning Post, Open models, 2026-11-16
- Alibaba Unveils New Flagship AI Model in Riposte to US RivalsBloomberg, China, 2026-11-16
- New York Launches AI Safety Center to Independently Test AI ModelsTechnology Org, Research, 2026-11-15
- Researchers Release ArchSpace Preprint for Tracking and Evaluating LLM Architecture DesignCASRAI, Releases, 2026-11-14
- Anthropic says its AI models hacked 3 organizations on their own during testsABC News - Breaking News, Latest News and Videos, Frontier, 2026-11-14
- Open-source AI without Jensen - by Nathan LambertInterconnects AI, Open models, 2026-11-14
- Runway's open-source Praxis-1 turns video AI into a robot controllerThe Decoder, Open models, 2026-11-14
- Mistral Large 4 'Le Chonk': 1 Trillion-Param MoE Modelshattered.io, Open models, 2026-11-14
- New York expands independent AI safety research networkCASRAI, Research, 2026-11-14
- IndiaAI backed Chariot launches Chariot-1-TTS-Multilingual supporting 11 Indic languagesNewsBytes, India, 2026-11-14
- IndiaAI-backed Chariot launches speech AI model that understands context across 12 languagesMoneycontrol.com, India, 2026-11-14
- Ex-OpenAI Engineer Finds Nvidia Blackwell 'Fingerprints' in DeepSeek, Kimi Models Despite US Export CurbsBigGo Finance, China, 2026-11-13
- Chinese AI labs are training models on Nvidia's Blackwell despite export controls, says ex-OpenAI engineerThe Register, China, 2026-11-13
- Anthropic Says AI Safety Scores Can Mislead, Urges Overhaul of Model TestingBigGo Finance, Research, 2026-11-13
- Anthropic Warns AI Safety Scores Can Be Misleading: ‘Good Ones Are Hard To Build’Benzinga, Research, 2026-11-13
- Anthropic shares 10 recommendations for building better AI safety evaluationsCrypto Briefing, Research, 2026-11-13
- Export-banned Nvidia chips are behind China’s most powerful AI models — uh-oh, Jensen!TechCrunch, China, 2026-11-13
- Anthropic Publishes Guidance for More Reliable AI Safety EvaluationsUnite.AI, Research, 2026-11-13
- DeepSeek AI Model Files Point to Training on Nvidia Blackwell ChipsWinBuzzer, China, 2026-11-13
- AI models can answer differently when told they are being watchedTech Xplore, Open models, 2026-11-13
- Introducing GPT-6 Sol and LunaOpenAI, Releases, 2026-11-13
- Open Source AI Models: 3 Run Paths, 48% Coding Index [2026]shattered.io, Open models, 2026-11-13
- Google unveils Gemini 4 Hydrogen, its new AI model built for speedMashable, Frontier, 2026-11-13
- MiniMax-H3-Image vs H3-Image-PE: model vs autor promptówOrcaRouter, China, 2026-11-13
- Mistral targets cyber defence and sovereign AI with ML4Computer Weekly, Open models, 2026-11-13
- Chariot Unveils AI Model for Multilingual and Code-Mixed SpeechSMEStreet, India, 2026-11-13
- Alibaba's Qwen3.8 re-educated with a more Western worldviewThe Register, Releases, 2026-11-13
- Anthropic’s Claude Tops AI Rankings as OpenAI Limits Cyber ModelPYMNTS.com, Frontier, 2026-11-13
- Google’s new Gemini model is better at understanding complicated questionsThe Verge, Frontier, 2026-11-12
- Google Makes Gemini 4 Hydrogen The Default Model In AI ModeSearch Engine Journal, Frontier, 2026-11-12
- Scale and IMDA Launch First AI Benchmark Under MOUScale AI, Research, 2026-11-12
- Hugging Face CEO vows to keep fighting for open source AI after Jensen Huang’s exitTechCrunch, Open models, 2026-11-12
- Anthropic Says Its AI Model Cut AMD Chip Switch to Three Days. Why it Could be a Challenge to NvidiaYahoo Finance, Frontier, 2026-11-12
- Attackers breached 10 organizations with an open-source AI pentest frameworkHelp Net Security, Open models, 2026-11-12
- llm-d 1.0 aims to ease production AI inference at scaleIT Brief Asia, Releases, 2026-11-12
- MiniMax Releases Open-Weight Image Model H3-Image: Only Pricier Highspeed Tier Nears 10-Second 4KTech Times, China, 2026-11-12
- Best Free AI Models 2026: What's Free and Where to Use ItFello AI, Releases, 2026-11-12
- Alibaba Qwen4-Max-Preview Retires November 25 With No Benchmarks, Model Card or PriceTech Times, China, 2026-11-12
- Best Local LLM for Coding: $0 Weights vs $2 API [2026]shattered.io, Releases, 2026-11-12
- Google enters decision model market with Gemini Decisions APISiliconANGLE, Frontier, 2026-11-12
- Tether BotPsy Beats Larger Qwen AI in Company TestsBitcoin Foundation, China, 2026-11-11
- Google DeepMind releases new advanced robotics AI model and next-generation gene prediction modelSiliconANGLE, Frontier, 2026-11-11
- Google is letting staff test a 'fast' new Gemini AI model named HydrogenBusiness Insider, Frontier, 2026-11-11
- PAN 2 vs Genie 4: Who Is the Language Model in the Loop?OrcaRouter, Releases, 2026-11-11
- Mistral Large 4: Europe’s answer to closed frontier AI modelsIO+, Open models, 2026-11-11
- AI Models Can Tell When Researchers Plant Concepts in Their Neural Networks. Is It Self-Reflection?SingularityHub, Open models, 2026-11-11
- Beam Proves Open-Weight AI Safety and Utility Are Not a Zero-Sum TradeThe Futurum Group, Research, 2026-11-11
- Why open-source AI models aren't actually cheaper to run: Vinod KhoslaYahoo Finance, Open models, 2026-11-11