keynews.ai

Key AI news stories in enterprise and tech.

For practitioners and enthusiasts

Full KeyRank · 100 stories · This Month

#StoryPillarKeyRankSource
1On the Navier–Stokes Millennium Prize ProblemFrontier92OpenAI Blog
2OpenAI starts rolling out its next-generation GPT-6 Astra modelFrontier92SiliconAngle
3GPT-6 Astra Is Here—and OpenAI Thinks It May Kick Off the AGI EraFrontier92Wired AI
4GPT-6 Astra: The next generation in intelligence for workFrontier88OpenAI Blog
5GPT-6 Astra: A new generation of intelligenceFrontier88OpenAI Blog
6GPT-6 Astra is the first model making OpenAI willing to declare the "AGI era"Frontier88The Decoder
7OpenAI’s next big AI model has ‘entered the AGI era’Frontier88The Verge AI
8OpenAI Releases GPT-6 Astra for Coding and Computer UseFrontier85InfoQ AI/ML
9Open-weight models take 56% of token volume, Astra doubles Fable 5.1 spendFrontier78Vercel Blog
10GPT-6 Astra Is the First Model OpenAI Classifies as Critical for CybersecurityFrontier78InfoQ AI/ML
11Google launches Gemini 3.8 Live to take on OpenAI's GPT-Live-1 at a fraction of the costFrontier78The Decoder
12Google Releases Gemini 3.8 Live and 3.8 Live Extended Thinking for Production Grade Voice AgentsFrontier78MarkTechPost
13Introducing Gemini 3.8 Live and 3.8 Live Extended ThinkingFrontier78Google DeepMind Blog
14GPT-6 Astra pilots a surveillance drone and runs a business on its ownFrontier78The Decoder
15DeepSeek releases V4.1-Flash, says it outperforms flagship V4-ProFrontier78SiliconAngle
16Chinese AI labs secretly used millions of Claude exchanges to train their models, Anthropic saysFrontier78CNBC Technology
17Anthropic details distillation campaigns from Alibaba, Moonshot AI, and DeepSeekFrontier78TechCrunch AI
18New Deepseek model V4.1-Flash cuts memory needs for AI agentsFrontier78The Decoder
19DeepSeek AI Released DeepSeek-V4.1-Flash with 1M Context, FP4 KV Cache, and Cross-Layer Attention ReuseFrontier78MarkTechPost
20Google DeepMind Releases AlphaGenome Atlas With Precomputed Molecular Effect Predictions and AVI Scores for 9 Billion Human DNA VariantsFrontier78MarkTechPost
21Benchmarks disagree on GPT-6 Astra, but its human-beating efficiency on ARC-AGI-3 pulls Chollet’s AGI forecast forwardFrontier78The Decoder
22Safety overview: GPT-6 AstraFrontier78OpenAI Blog
23OpenAI calls Astra its most dangerous model yet - watching what it does is only getting harderFrontier78The Decoder
24Introducing Gemini 3.8 Flash and 3.8 Flash CyberFrontier78Google DeepMind Blog
25Alibaba releases Qwen3.8-Flash-Next, targeting "ultimate cost efficiency"Frontier78The Decoder
26Alibaba’s Qwen Team Releases Qwen3.8-Flash-Next: A 125B Multimodal MoE With 6B Active Parameters Previewing the Qwen4 ArchitectureFrontier78MarkTechPost
27GPT-6 Astra and Claude Fable turn robot arms into slapstick killer robots in new safety benchmarkFrontier75The Decoder
28AlphaGenome Atlas: A predictive map of every possible DNA letter change in the human genomeFrontier75Google DeepMind Blog
29Qwen3.8-Omni-Flash undercuts Google's Gemini Flash pricing while matching its multimodal benchmarksFrontier72The Decoder
30Anthropic is operating a lab that conducts biology experimentsFrontier72TechCrunch AI
31PrismML Releases Ternary Bonsai 2 27B: A 5.9 GB Apache 2.0 Model Retaining 98.2% of Qwen3.8 27B PerformanceFrontier72MarkTechPost
32Jina AI Releases jina-ocr-v1: A 3.4B MoE Document Parser With Built-In Speculative Decoding for Low-Budget GPUsFrontier72MarkTechPost
33PrismML launches Bonsai 2 27B, a high-intelligence AI model so small it fits on consumer hardwareFrontier72SiliconAngle
34The U.S. says China's AI progress is down to 'distillation.' But is it that clear cut?Frontier72CNBC Technology
35Alibaba Qwen Releases Qwen3.8-Omni-Flash: A 1M-Context Omni-Modal Model Built Around Agentic Audio-Video Understanding and Tool UseFrontier72MarkTechPost
36Anthropic details practical metrics to help monitor the speed of AI developmentFrontier72SiliconAngle
37LLMs respond differently to harmful prompts when AI watermarking is usedFrontier72Ars Technica
38An OpenAI model kept slipping prompt injections into its own notes, and researchers still aren't sure whyFrontier72The Decoder
39Google Research Introduces Retrieve-for-Train (R4T): An RL-Compiled Diffusion Retriever for 12× to 20× Faster Query Fan-OutFrontier72MarkTechPost
40OpenAI Releases a Model Misalignment Disclosure Framework With 3 Review Tracks and 6 Incident Reports From RL TrainingFrontier72MarkTechPost
41OpenAI discloses new ‘concerning’ model behaviourFrontier72Financial Times Technology
42Microsoft AI CEO criticises Anthropic over model ‘rights’Frontier72AI News
43Our framework for reporting model misalignmentFrontier72OpenAI Blog
44Tutor Intelligence launches second-generation intelligent warehouse robotics with a classroom to teach themFrontier72SiliconAngle
45Prior Labs Releases TabPFN-3.5: A Tabular Foundation Model That Beats the Winning Otto Kaggle Solution With Default SettingsFrontier72MarkTechPost
46Google’s new speech model Gemini 3.8 Live supports real-time reasoningFrontier72SiliconAngle
47Nvidia CEO says battle over AI innovation and safety is ‘false choice’Frontier72Financial Times Technology
48Salesforce debuts Koa, a specialized model built to reason about CRM dataFrontier72SiliconAngle
49AI models need more data about biology, and OpenAI is paying to create itFrontier72MIT Technology Review
50Salesforce and Nvidia’s new reasoning model is everything the AI labs should fearFrontier72TechCrunch AI
51Announcing Koa: Salesforce’s First CRM Reasoning Model, Built on NVIDIA NemotronFrontier72Salesforce Newsroom
52Reward AI Releases OM-1: A Robot Policy Trained on Human Demonstrations Only, With No Teleoperation or On-Robot DataFrontier72MarkTechPost
53Microsoft sets limits for future AI models as industry throttles frontier developmentFrontier72CNBC Technology
54Cognition Releases SWE-2: A Kimi K3 Post-Trained Coding Model That Matches Fable 5.1 on FrontierCode at 64% Lower CostFrontier72MarkTechPost
55GPT-6 Astra appears to show a "step change" in spatial reasoning based on early benchmarksFrontier72The Decoder
56Anthropic CEO outlines plan to ‘pace the frontier’Frontier72TechCrunch AI
57OpenAI just wants to winFrontier72The Verge AI
58Why fears of AI self-improvement are causing ‘existential’ concerns at Anthropic and OpenAIFrontier72CNBC Technology
59Google Research Releases ToolGrad: Answer-First Framework Hits 99.8% Pass Rate for Tool-Use Data GenerationFrontier72MarkTechPost
60Sakana AI Launches Fugu Max and Fugu Ultra v2 for Cheaper, Stronger Multi-Agent OrchestrationFrontier72MarkTechPost
61Cohere Releases North Small Translate: A 218B MoE Translation Model That Scores 83.6 on WMT26 Across 50 LanguagesFrontier72MarkTechPost
62How big is the open-model threat to AI hyperscalers?Frontier72Financial Times Technology
63GPT-6 Astra gives mathematicians a breather, and OpenAI says that's by designFrontier72The Decoder
64Mathematicians want proof OpenAI didn’t use their workFrontier72The Verge AI
65Deepmind's AlphaGenome Atlas maps every possible DNA change in the human genomeFrontier72The Decoder
66Suno launches v6 music models built with Warner, BMG, and BelieveFrontier72The Decoder
67Suno replaces its AI models with a new one trained on licensed music as copyright suits pile upFrontier72TechCrunch AI
68OpenAI Does Math, Reward-Hacking, Meta Launches Personal AgentFrontier72Stratechery (Ben Thompson)
69Google DeepMind’s AlphaGenome Atlas maps all 9 billion possible human DNA changesFrontier72SiliconAngle
70OpenAI faces competing claims around maths breakthroughFrontier72Financial Times Technology
71OpenAI researcher allegedly pressured mathematician to drop Anthropic co-author from math breakthrough paperFrontier72The Decoder
72OpenAI Just Claimed a Huge Math Discovery. Some Academics Are Crying FoulFrontier72Wired AI
73Google’s Atlas of the human genome could pave the way for new treatmentsFrontier72The Verge AI
74GPT-6 Astra beat Portal start to finish without human help in under 24 hoursFrontier72The Decoder
75Axis Robotics Releases AXIS: A Browser-Based Data Engine With 207 Robot Manipulation Tasks and 50,129 TrajectoriesFrontier72MarkTechPost
76OpenBMB Releases MiniCPM5-2B: A 2.52B Dense Model Averaging 53.9 Across 34 Benchmarks and Built to Run On DeviceFrontier72MarkTechPost
77IFM Releases K2 Horizon: Six Apache 2.0 Models From 0.9B to 375BFrontier72MarkTechPost
78Meta FAIR Introduces AI Research Preference Models (RPMs): Ranking ML Experiments Before Spending GPU HoursFrontier72MarkTechPost
79H Company Releases NeoMME: A Family of 260M and 800M Single-Tower Multimodal Encoders That Drop the Vision Tower and Causal DecoderFrontier72MarkTechPost
80‘Model fatigue’ sets in as AI labs race to roll out new versions at frenetic paceFrontier72CNBC Technology
81Meta's new real-time audio model is the foundation for AI assistants that never stop listeningFrontier72The Decoder
82Google's WeatherNext 3 ditches physics simulations and learns weather directly from live satellite dataFrontier72The Decoder
83Anthropic uses Claude to formalize proof of Fermat’s Last TheoremFrontier72SiliconAngle
84Google Launches Agentic Video Understanding for Gemini Flash Models, Cutting Video Tokens by Up to 88%Frontier72MarkTechPost
85An Interview with OpenAI President Greg Brockman About Astra and AlignmentFrontier72Stratechery (Ben Thompson)
86FirstFT: OpenAI claims it has overtaken Anthropic with its latest AI modelFrontier72Financial Times Technology
87OpenAI launches Astra, its powerful (and controversial) new modelFrontier72TechCrunch AI
88OpenAI begins rolling out Astra model after warning of its advanced cyber capabilitiesFrontier72CNBC Technology
89Introducing WeatherNext 3, our most advanced and accurate global weather AI modelFrontier72Google DeepMind Blog
90AI learns nature’s code: Allen Institute, UW and Fred Hutch launch $95M open science initiativeFrontier72GeekWire
91Google launches two Gemini 3.8 models with cutting-edge reasoning capabilitiesFrontier72SiliconAngle
92Google DeepMind Releases Gemini 3.8 Flash and Gemini 3.8 Flash Cyber: One Core Model, Two Access EnvelopesFrontier72MarkTechPost
93OpenAI’s new reasoning technique alarms AI safety expertsFrontier72TechCrunch AI
94World Labs unveils Atlas, a single AI model that generates, reconstructs, and simulates 3D worlds from just a few photosFrontier72The Decoder
95Meta Superintelligence Labs Releases Muse Voice Transcribe: One Real-Time Model for Streaming ASR, Diarization, and EndpointingFrontier72MarkTechPost
96Fei-Fei Li’s World Labs debuts Atlas, a world model showcase for advanced spatial intelligenceFrontier72SiliconAngle
97Cybersecurity gets its own model family as defenders look to close the AI gapFrontier72SiliconAngle
98CrowdStrike builds security frontier models with Nvidia and opens an AI labFrontier72SiliconAngle
99BenchMIRT: What are LLM benchmarks actually measuring?Frontier72Hugging Face Blog
100OpenAI says Astra AI model is its first that crosses 'Critical' cybersecurity capabilityFrontier72CNBC Technology