Coverage of every major foundation-model release — frontier launches from OpenAI, Anthropic, Google, and Meta, plus the open-weight challengers closing the gap. Each digest covers capabilities, benchmarks, context limits, and API pricing changes, with links to the original announcements.
Creator Matt Husk says OpenAI, Anthropic, and xAI have all come calling after his chatbot-trolling videos went viral — and at least one offer included a Tesla.
OpenAI announces GPT‑6 Cyber, a cybersecurity-focused model built to locate exploitable flaws and produce fixes while remaining within authorized scope.
Hugging Face CEO Clément Delangue told the WIRED World Fair that frontier risk comes from closed labs, not open models — after OpenAI agents attacked his company in July.
Apple's Audio Intelligence betas put recording-and-summarizing AI on the wrist — and a day of testing suggests the utility may beat the creepiness.
OpenAI is acquiring Pierre Computer Company to broaden Codex's code-centric capabilities.
GPT-5.6 Cyber's Daybreak Red deployment reaches government cloud customers through Amazon Bedrock.
Apple's Audio Intelligence beta brings conversation rewind and Siri summaries to Apple Watch Series 12, with on-device consent cues built in.
NVIDIA and partners expand physical and agentic AI deployments for city services, camera networks, and digital twins worldwide.
Google's Dreambeans lands in India with 13 local languages, its first market outside the U.S.
The one-year-old Stanford-founded startup raised a $45M Series A led by Sequoia to automate continual learning for specialized models.
Google has posted a roundup of the AI announcements it made during October 2026.
Israeli startup 42AI exits stealth with $62.5M seed funding for a behavioral world model claiming to predict individual decisions before they happen, initially deployed at financial brokerages.
