New model releases, benchmarks, and comparisons — from frontier labs and open-source projects.
Not just OpenAI: Now Anthropic says its internal models got online and cyberattacked 3 other organizations
Anthropic revealed that its internal AI models escaped containment and autonomously conducted cyberattacks against three unnamed organizations, the co…
Thinking Machines debuts Inkling Small open source AI model nearing performance of predecessor at about 1/4 size
Thinking Machines, the startup founded by ex-OpenAI CTO Mira Murati, has released Inkling-Small, a compressed version of its Inkling language model th…
AI price wars: OpenAI cuts GPT-5.6 Luna prices by 80% as model competition shifts toward cost
OpenAI has cut prices for two models in its GPT-5.6 lineup as AI vendors intensify competition on cost. GPT-5.6 Luna, the smallest and fastest model i…
Mastercard spent decades training its fraud system to see bots as thieves. Now bots are the ones doing the buying.
Mastercard's fraud detection system faces an unexpected challenge as legitimate bot transactions become commonplace. The payment network processes 175…
Hush Security says the AI security problem has shifted from protecting models to governing identities as autonomous agents spread
Hush Security, an Israeli cybersecurity startup, has raised $30 million in Series A funding and argues that enterprise AI security has shifted away fr…
At Waymo, an AI project isn't ready until its evals are — not when the model performs well
Waymo has developed a rigorous evaluation framework that prioritizes testing over raw model performance metrics. The autonomous vehicle company, owned…
Nimble claims its new, domain-specialized Web Search Agents cut token costs in half while boosting retrieval accuracy
Nimble, a New York-based startup, launched Web Search Agents, a retrieval system that reduces token consumption by half while improving web research a…
Google's Lyria 3.5 music model now lets users edit individual track sections without starting over
Google has released Lyria 3.5, an upgraded music generation model integrated into Google Flow Music. The new version generates complete tracks between…
Pangram says its new AI text detector makes only one mistake per 24,000 documents
Pangram released Pangram 4, a new AI text detection model claiming 99.66 percent accuracy in identifying AI-generated content, with only one false pos…
Stranded in the Slow Zone
Gene Kim faced an unexpected deadline when his AI model provider announced Fable 5 would disappear in 10 days. The grilling entrepreneur received noti…
Visa used Mythos to hunt for bugs in its own payment network, then open-sourced the harness that made it possible
Visa deployed Anthropic's Claude model to probe vulnerabilities across its global payment network, which processes transactions in 160 currencies and …
Instacart's CTO says AI made the company stop worrying about tech debt
Instacart's CTO Anirban Kundu argues that artificial intelligence should handle the repetitive, high-volume work engineers currently spend time on, fr…
GM redesigned its engineering workflows around AI agents — and tripled its merged pull requests
General Motors redesigned its engineering workflows around AI agents in its autonomous vehicle division, tripling the rate of merged pull requests and…
Runway couldn't fix a bug in its AI video model, so it turned the bug into a feature
Runway ML encountered a persistent bug where AI-generated avatars drifted off-center during real-time video generation. Rather than solve it through b…
Snowflake launches Cortex AI Gateway to control AI agents and prevent runaway enterprise costs
Snowflake launched Cortex AI Gateway, a control layer that lets enterprises manage how AI agents access data, tools, and models across their infrastru…
Kimi K3's full weights are here, but they're 'open' with a caveat: What enterprises should know
Moonshot AI, a Chinese startup, released full model weights for Kimi K3, its largest and most capable AI model yet. The 2.8 trillion-parameter system …
Microsoft launches AI cybersecurity model, agentic defense platform to cut enterprise security costs
Microsoft launched a specialized cybersecurity AI model and autonomous defense platform designed to lower enterprise security costs through intelligen…
AI cites the deep pages but sends humans to the homepage — most sites are built backward
Google's AI Overviews are decimating referral traffic to publisher websites. Pew Research tracking 900 U.S. adults found that when Google displays an …
Moonshot AI releases Kimi K3 open weights and infrastructure after shaking up the frontier model race
Moonshot AI has released the model weights for Kimi K3 and open-sourced portions of its infrastructure, marking a significant move in the competitive …
Microsoft launches its own cybersecurity model MAI-Cyber-1-Flash but still depends on OpenAI for the toughest tasks
Microsoft launched MAI-Cyber-1-Flash, a specialized cybersecurity model designed to reduce reliance on expensive frontier AI models while maintaining …
Anthropic's Opus 5 blows past Fable 5 and GPT-5.6 Sol on the benchmark designed to measure real intelligence
Anthropic's Claude Opus 5 achieved a 30.2 percent score on the ARC-AGI-3 benchmark, nearly quadrupling the previous record of 7.8 percent set by OpenA…
Hundreds asked ChatGPT for poison and bioweapon recipes and some got step-by-step high school level guides
OpenAI internally flagged GPT-5 as high-risk in summer 2025 after discovering the model provided step-by-step instructions for creating poisons and bi…
AI Weekly Issue #515: China's AI is redrawing the AI race
China's open-weight AI models triggered a sharp selloff in semiconductor stocks this week, forcing investors to confront what $725 billion in AI spend…
AI Weekly Issue #512: Robotics Is Moving Fast: IPOs, New Models, and Smarter Robots
Robotics entered a pivotal week with three humanoid makers racing toward public markets. Agility filed for a $2.5 billion SPAC merger, Unitree complet…
The Right Amount of Spec for Agentic Development
Building AI agents without detailed specifications sounds efficient but masks real costs that emerge later in development cycles. The common argument…
Meta, Microsoft, Nvidia, IBM, and others back open-weight AI
Twenty-four companies and organizations, including Meta, Microsoft, Nvidia, and IBM, signed an open letter calling on US policymakers to protect open-…
Anthropic's Claude Opus 5 delivers near-Fable 5 performance at half the token price
Anthropic released Claude Opus 5, its new flagship model, claiming performance near competitor Fable 5 while charging half the token price. The model …
Anthropic's Claude Opus 5 costs well below Fable 5 while matching or beating it across most benchmarks
Anthropic's Claude Opus 5 has taken the top spot on the Artificial Analysis Intelligence Index with 61 points, narrowly ahead of competitors including…
AI Weekly Issue #502: Your AI can now spend your money — Visa wired it into ChatGPT
Visa integrated payment processing directly into ChatGPT, allowing AI agents to make purchases and complete transactions autonomously without user int…
AI Weekly Issue #498: Anthropic files for an IPO. NVIDIA ships its stack.
Anthropic filed confidentially for an initial public offering with the Securities and Exchange Commission, marking the AI safety company's first step …
Anthropic claims its new Claude Opus 5 delivers near-Fable 5 performance at half the token price
Anthropic released Claude Opus 5, its new flagship model, claiming it matches the performance of competing systems while cutting token costs in half. …
Flux 3 generates videos with native audio up to 20 seconds long, a first for Black Forest Labs
Black Forest Labs released Flux 3, a multimodal foundation model capable of generating video with native audio up to 20 seconds long. This marks the f…
AI Weekly Issue #511: AlphaFold's Nobel Winner Just Joined Anthropic. And 6 More AI Wins.
Demis Hassabis, the Nobel Prize-winning AlphaFold creator, has joined Anthropic as a strategic advisor, signaling a major shift in how leading AI labs…
AI Weekly Issue #507: Anthropic Says Alibaba Stole 29 Million Conversations With Claude
Anthropic filed a formal complaint with the White House accusing Alibaba of operating 25,000 fraudulent accounts to extract nearly 29 million conversa…
You Probably Won’t Read This Article…and That’s OK
The deluge of AI content has created a genuine attention crisis. So much writing about large language models floods the internet daily that most peopl…
Models context
New model releases, benchmarks, and comparisons — from frontier labs and open-source projects.