| 1 | On the Navier–Stokes Millennium Prize Problem | Frontier | 92 | OpenAI Blog |
| 2 | OpenAI starts rolling out its next-generation GPT-6 Astra model | Frontier | 92 | SiliconAngle |
| 3 | GPT-6 Astra Is Here—and OpenAI Thinks It May Kick Off the AGI Era | Frontier | 92 | Wired AI |
| 4 | Claude Opus 5.5, GPT-6 Sol, GPT-6 Luna, and a new price war | Frontier | 88 | Simon Willison |
| 5 | Introducing GPT-6 Sol and Luna | Frontier | 88 | OpenAI Blog |
| 6 | GPT-6 Astra: The next generation in intelligence for workRising | Frontier | 88 | OpenAI Blog |
| 7 | OpenAI’s next big AI model has ‘entered the AGI era’ | Frontier | 88 | The Verge AI |
| 8 | OpenAI Releases GPT-6 Astra for Coding and Computer Use | Frontier | 85 | InfoQ AI/ML |
| 9 | GPT-6 Astra is the first model making OpenAI willing to declare the "AGI era"Rising | Frontier | 88 | The Decoder |
| 10 | DeepSeek releases V4.1-Flash, says it outperforms flagship V4-ProRising | Frontier | 78 | SiliconAngle |
| 11 | GPT-6 Astra Is the First Model OpenAI Classifies as Critical for CybersecurityRising | Frontier | 78 | InfoQ AI/ML |
| 12 | Introducing Gemini 3.8 Live and 3.8 Live Extended ThinkingRising | Frontier | 78 | Google DeepMind Blog |
| 13 | Salesforce and Nvidia’s new reasoning model is everything the AI labs should fearRising | Frontier | 72 | TechCrunch AI |
| 14 | How big is the open-model threat to AI hyperscalers?Rising | Frontier | 72 | Financial Times Technology |
| 15 | Anthropic’s biolab made a discovery it’s comparing to Crispr | Frontier | 78 | The Verge AI |
| 16 | Anthropic Releases Claude Opus 5.5: Fable 5.1-Level Performance at 40% Lower Running Cost Than Opus 5 | Frontier | 78 | MarkTechPost |
| 17 | Anthropic and OpenAI roll out cheaper models in first release since call for slowdown | Frontier | 78 | CNBC Technology |
| 18 | Xiaomi's affordable flagship AI leads the open models, and Anthropic says Claude helped get it there | Frontier | 78 | The Decoder |
| 19 | Open-weight models take 56% of token volume, Astra doubles Fable 5.1 spend | Frontier | 78 | Vercel Blog |
| 20 | Google launches Gemini 3.8 Live to take on OpenAI's GPT-Live-1 at a fraction of the cost | Frontier | 78 | The Decoder |
| 21 | Google Releases Gemini 3.8 Live and 3.8 Live Extended Thinking for Production Grade Voice Agents | Frontier | 78 | MarkTechPost |
| 22 | Chinese AI labs secretly used millions of Claude exchanges to train their models, Anthropic says | Frontier | 78 | CNBC Technology |
| 23 | Anthropic details distillation campaigns from Alibaba, Moonshot AI, and DeepSeek | Frontier | 78 | TechCrunch AI |
| 24 | New Deepseek model V4.1-Flash cuts memory needs for AI agents | Frontier | 78 | The Decoder |
| 25 | DeepSeek AI Released DeepSeek-V4.1-Flash with 1M Context, FP4 KV Cache, and Cross-Layer Attention Reuse | Frontier | 78 | MarkTechPost |
| 26 | Google DeepMind Releases AlphaGenome Atlas With Precomputed Molecular Effect Predictions and AVI Scores for 9 Billion Human DNA Variants | Frontier | 78 | MarkTechPost |
| 27 | Benchmarks disagree on GPT-6 Astra, but its human-beating efficiency on ARC-AGI-3 pulls Chollet’s AGI forecast forward | Frontier | 78 | The Decoder |
| 28 | Safety overview: GPT-6 Astra | Frontier | 78 | OpenAI Blog |
| 29 | OpenAI calls Astra its most dangerous model yet - watching what it does is only getting harder | Frontier | 78 | The Decoder |
| 30 | Introducing Gemini 3.8 Flash and 3.8 Flash Cyber | Frontier | 78 | Google DeepMind Blog |
| 31 | Alibaba releases Qwen3.8-Flash-Next, targeting "ultimate cost efficiency" | Frontier | 78 | The Decoder |
| 32 | Alibaba’s Qwen Team Releases Qwen3.8-Flash-Next: A 125B Multimodal MoE With 6B Active Parameters Previewing the Qwen4 Architecture | Frontier | 78 | MarkTechPost |
| 33 | GPT-6 Astra and Claude Fable turn robot arms into slapstick killer robots in new safety benchmarkRising | Frontier | 75 | The Decoder |
| 34 | GPT-6 Astra pilots a surveillance drone and runs a business on its ownRising | Frontier | 78 | The Decoder |
| 35 | Our framework for reporting model misalignmentRising | Frontier | 72 | OpenAI Blog |
| 36 | Salesforce debuts Koa, a specialized model built to reason about CRM dataRising | Frontier | 72 | SiliconAngle |
| 37 | SpaceX launches Grok 4.7 with long-horizon processing, safety upgrades | Frontier | 75 | SiliconAngle |
| 38 | AlphaGenome Atlas: A predictive map of every possible DNA letter change in the human genome | Frontier | 75 | Google DeepMind Blog |
| 39 | Tutor Intelligence launches second-generation intelligent warehouse robotics with a classroom to teach themRising | Frontier | 72 | SiliconAngle |
| 40 | Google Research Introduces Retrieve-for-Train (R4T): An RL-Compiled Diffusion Retriever for 12× to 20× Faster Query Fan-OutRising | Frontier | 72 | MarkTechPost |
| 41 | Sakana AI Launches Fugu Max and Fugu Ultra v2 for Cheaper, Stronger Multi-Agent OrchestrationRising | Frontier | 72 | MarkTechPost |
| 42 | Deepmind was built to chase AGI, but its new chief just wants Gemini 4 out the door | Frontier | 72 | The Decoder |
| 43 | Anthropic says Claude discovered a new enzyme system, but CRISPR researchers call it routine genome mining | Frontier | 72 | The Decoder |
| 44 | Black Forest Labs launches FLUX 3 Action, an open robotics AI model | Frontier | 72 | The Decoder |
| 45 | Google launches two benchmark-topping speech generation models | Frontier | 72 | SiliconAngle |
| 46 | NVIDIA Releases Nemotron 3 Diarization: A 100M-Parameter Open-Weight Model That Tracks 8 Speakers in Real Time | Frontier | 72 | MarkTechPost |
| 47 | Introducing MentalHealthBench | Frontier | 72 | OpenAI Blog |
| 48 | Anthropic unveils Opus 5.5: powerful performance, still premium price | Frontier | 72 | AI Business |
| 49 | Alibaba launches Qwen Audio 3.1 with new models and slashes AI audio prices by up to 95 percent | Frontier | 72 | The Decoder |
| 50 | Alibaba Unveils Zhenwu V900 — and Plans Qwen Models With Up to 10 Trillion Parameters | Frontier | 72 | TechRepublic |
| 51 | Kyutai Releases Voice of Reason: A Speech-Native Model that Solves Spoken Math with Reinforcement Learning | Frontier | 72 | MarkTechPost |
| 52 | Meta admits Muse’s likeness to OpenClaw isn’t a coincidence | Frontier | 72 | TechCrunch AI |
| 53 | Google’s robotics unit Intrinsic open-sources its foundational infrastructure for intelligent robots | Frontier | 72 | SiliconAngle |
| 54 | OpenAI says its internal model solved over 100 long-standing math problems after just a month of training | Frontier | 72 | The Decoder |
| 55 | Anthropic Builds Biology Lab to Test What Claude Can Do in the Real World | Frontier | 72 | TechRepublic |
| 56 | Jev introduces a new shape of LLM - System One, aka Decision Models | Frontier | 72 | Simon Willison |
| 57 | OpenAI forms math advisory group as its AI resolves more than 100 open problems | Frontier | 72 | TechCrunch AI |
| 58 | Alibaba Qwen Releases Qwen-Image-2.1: A 7B Open-Weight Model for Image Generation and Editing | Frontier | 72 | MarkTechPost |
| 59 | StepFun Launches Step 5 Preview: A 600B-Total, 27B-Active MoE Model With 1M Context for Long-Horizon Agentic Work | Frontier | 72 | MarkTechPost |
| 60 | TypeSafe AI’s new models work with machines, not humans | Frontier | 72 | CIO |
| 61 | Tether addresses AI underinvestment in Africa with open-source machine translation models | Frontier | 72 | CIO |
| 62 | Anthropic is operating a lab that conducts biology experiments | Frontier | 72 | TechCrunch AI |
| 63 | PrismML Releases Ternary Bonsai 2 27B: A 5.9 GB Apache 2.0 Model Retaining 98.2% of Qwen3.8 27B Performance | Frontier | 72 | MarkTechPost |
| 64 | Jina AI Releases jina-ocr-v1: A 3.4B MoE Document Parser With Built-In Speculative Decoding for Low-Budget GPUs | Frontier | 72 | MarkTechPost |
| 65 | PrismML launches Bonsai 2 27B, a high-intelligence AI model so small it fits on consumer hardware | Frontier | 72 | SiliconAngle |
| 66 | The U.S. says China's AI progress is down to 'distillation.' But is it that clear cut? | Frontier | 72 | CNBC Technology |
| 67 | Alibaba Qwen Releases Qwen3.8-Omni-Flash: A 1M-Context Omni-Modal Model Built Around Agentic Audio-Video Understanding and Tool Use | Frontier | 72 | MarkTechPost |
| 68 | Anthropic details practical metrics to help monitor the speed of AI development | Frontier | 72 | SiliconAngle |
| 69 | LLMs respond differently to harmful prompts when AI watermarking is used | Frontier | 72 | Ars Technica |
| 70 | An OpenAI model kept slipping prompt injections into its own notes, and researchers still aren't sure why | Frontier | 72 | The Decoder |
| 71 | Microsoft AI CEO criticises Anthropic over model ‘rights’ | Frontier | 72 | AI News |
| 72 | Prior Labs Releases TabPFN-3.5: A Tabular Foundation Model That Beats the Winning Otto Kaggle Solution With Default SettingsRising | Frontier | 72 | MarkTechPost |
| 73 | Google’s new speech model Gemini 3.8 Live supports real-time reasoning | Frontier | 72 | SiliconAngle |
| 74 | Nvidia CEO says battle over AI innovation and safety is ‘false choice’ | Frontier | 72 | Financial Times Technology |
| 75 | AI models need more data about biology, and OpenAI is paying to create it | Frontier | 72 | MIT Technology Review |
| 76 | Reward AI Releases OM-1: A Robot Policy Trained on Human Demonstrations Only, With No Teleoperation or On-Robot Data | Frontier | 72 | MarkTechPost |
| 77 | Microsoft sets limits for future AI models as industry throttles frontier development | Frontier | 72 | CNBC Technology |
| 78 | Cognition Releases SWE-2: A Kimi K3 Post-Trained Coding Model That Matches Fable 5.1 on FrontierCode at 64% Lower Cost | Frontier | 72 | MarkTechPost |
| 79 | GPT-6 Astra appears to show a "step change" in spatial reasoning based on early benchmarks | Frontier | 72 | The Decoder |
| 80 | Anthropic CEO outlines plan to ‘pace the frontier’ | Frontier | 72 | TechCrunch AI |
| 81 | OpenAI just wants to win | Frontier | 72 | The Verge AI |
| 82 | Google Research Releases ToolGrad: Answer-First Framework Hits 99.8% Pass Rate for Tool-Use Data Generation | Frontier | 72 | MarkTechPost |
| 83 | GPT-6 Astra gives mathematicians a breather, and OpenAI says that's by design | Frontier | 72 | The Decoder |
| 84 | Qwen3.8-Omni-Flash undercuts Google's Gemini Flash pricing while matching its multimodal benchmarksRising | Frontier | 72 | The Decoder |
| 85 | OpenAI Releases a Model Misalignment Disclosure Framework With 3 Review Tracks and 6 Incident Reports From RL TrainingRising | Frontier | 72 | MarkTechPost |
| 86 | OpenAI discloses new ‘concerning’ model behaviourRising | Frontier | 72 | Financial Times Technology |
| 87 | Announcing Koa: Salesforce’s First CRM Reasoning Model, Built on NVIDIA NemotronRising | Frontier | 72 | Salesforce Newsroom |
| 88 | Why fears of AI self-improvement are causing ‘existential’ concerns at Anthropic and OpenAIRising | Frontier | 72 | CNBC Technology |
| 89 | Cohere Releases North Small Translate: A 218B MoE Translation Model That Scores 83.6 on WMT26 Across 50 Languages | Frontier | 72 | MarkTechPost |