ChatGPT Wallet brings approved AI purchases to US Pro users
ChatGPT Wallet lets US Pro users approve AI purchases. See how payment tokens work, which cards qualify, and what happens to shopping data.
News about artificial intelligence: new models and tools, the companies that build them, research findings, and the rules that govern their use.
ChatGPT Wallet lets US Pro users approve AI purchases. See how payment tokens work, which cards qualify, and what happens to shopping data.
Smart glasses from Samsung and Google launch December 10, with Gemini, two frame collections and different rules for camera images and chats.
Gemini 4 Hydrogen reaches Search, apps and coding tools. Free access has limits, while Google’s model card explains what long uploads can do.
The Gemini CarPlay app brings voice chats to iPhone drivers. Learn how setup works, how Siri differs, and which personal features have limits.
Businesses in ChatGPT lets firms manage profiles and launch sales agents. OpenAI sets rules for labels, personal data and customer consent.
Anthropic IPO plans offer 825 million shares at $70–$78. Here is how the proceeds split and what the proposed Nasdaq listing means.
Google TPU 8tx adds private AI processing with AMD hosts, while 8i Max targets larger context. See how their designs and limits differ.
Claude Fable 5.2 and Mythos 5.2 bring new coding and research claims. Explore prices, nine benchmark charts, four demos and safety limits.
CheMeleon helps predict molecular traits with less task-specific data. Learn how it works, where it struggles and why lab safety tests still matter.
BioBasicsBench tests how AI reads real lab evidence. OpenAI reports gains, while errors show why context and ongoing checks still matter.
GPT-6 Cyber opens to approved security researchers. OpenAI reports stronger tests, while outside evidence shows why human oversight still matters.
Gemini Omni 1.2 Flash adds 15-second videos and lower output prices. Explore editing tools, app access and the limits that differ across Google services.
Nexus Mods AI rules require upload disclosures and detailed labels, with a January deadline for existing mods and easier content filters planned.
Anthropic adds Claude game cloning guidance to Claude Code on Opus 5.5, while saying original games and users’ own designs should still work.
OpenAI teen safety changes will restrict more weapons information. A first Community Safety Report is planned for early 2027, then every six months.