Google has released its third Google Home app update this September, focused on making Gemini for Home better at filtering background noise to register voice commands and respond faster. The update also improves multilingual recognition, Find My Phone reliability, smart light controls, camera tiles, and device setup, alongside fixes for Matter switches, Favorites tiles, camera playback, Onn camera history, and Automations.
Alphabet introduced Gemini 4 Argon, describing it as its most capable AI model to date, with gains in software engineering, cybersecurity and professional tasks like finance and legal work. Google says the model is already reducing memory needs at its data centers and being used by quantum computing researchers, and it will first go to trusted cybersecurity partners and the U.S. government for safety testing before wider release.
Artificial Analysis benchmarked Gemini 4 Argon (High), a model that accepts text and image input and outputs text with a 1M token context window. It scored 53 on the firm's Intelligence Index, well above the median of 26, while using 110M tokens during evaluation versus a median of 82M. Pricing is set at $2.00 per 1M input tokens and $10.00 per 1M output tokens, both matching the median for comparable models.
Google announced Gemini 4 Argon, a new frontier AI model it claims leads the industry in coding, knowledge work, and cybersecurity tasks. The model is not yet publicly available, though Google engineers are already using it internally, including for data center memory optimization and migrating large C/C++ codebases to Rust. Google released benchmark results showing Argon outperforming rivals like GPT-6 Astra, Fable 5.1, and Opus 5.5, and confirmed it will support a 1 million token output limit, up from 64,000 in earlier Gemini models.
The Trump administration unveiled America.gov, a free chatbot meant to consolidate roughly 29,000 government websites into a single tool answering questions about tasks like applying for a passport, registering to vote, or tracking a tax refund. The bot, powered by Google's Gemini and xAI's Grok, currently only provides guidance and directs users to relevant forms and sites rather than processing personal information, though officials say it will handle form submissions and applications like Medicare enrollment next year. Users quickly tested the chatbot and found it gave answers—such as confirming Joe Biden won the 2020 election and that climate change is real—that contradict statements made by Trump and his administration.
Google has unveiled Gemini 4 Argon, a new frontier AI model designed for sustained reasoning over complex, long-running tasks including software engineering, legal and finance work, and cybersecurity defense. The model is initially rolling out only to trusted testers through Google's Fairwind Program, alongside engagement with a U.S. government voluntary pre-release access process. Pricing has been set at $2 per million input tokens and $10 per million output tokens, with cached inputs discounted 95%.
Google published a phased schedule for replacing Gemini Gems with a new feature called Skills across Workspace and the Gemini app. Skills rollout begins October 2026, Gems move to a Settings panel by November 2026, and full removal of Gems is set for March 2027 for business/enterprise users and June 2027 for education accounts, with remaining Gems auto-migrated to draft skills.
Google introduced a new frontier AI model, Gemini 4 Argon, which DeepMind SVP Koray Kavukcuoglu says achieves strong results in software engineering, enterprise legal and finance tasks, and cybersecurity defense. The company is initially limiting access to select trusted cyber-defense partners while it engages with a U.S. government voluntary pre-release review process and strengthens safety measures against misuse and prompt injection before wider release.
Google has enrolled about 100 publishers in a pilot that pays them when their content contributes to Gemini-powered AI Overviews, according to The Information. Most participating small and mid-sized publishers reportedly receive payments equal to roughly one-tenth of one percent of their ad revenue, though one early participant is on pace to earn over $1 million annually while another gets $50,000-$60,000.
Google is reportedly running a test program, launched less than a year ago, that compensates roughly 100 publishers based on how much their content contributes to AI Overviews, AI Mode, and the Gemini chatbot, according to The Information via Digiday. One early participant reportedly earned over $1 million in a year, while a more recent entrant has earned $50,000 to $60,000.
An Android Authority writer describes using Anthropic's Claude as a logistics organizer rather than a decision-maker while planning a three-week trip to Japan. Instead of asking the AI to suggest destinations, he fed it his own research, preferences and must-see list, letting Claude arrange the details into a coherent schedule. He reports this method saved significant time compared with juggling Google Maps, spreadsheets and blog posts.
A user on Bluesky found that typing 'Play Minecraft' into the US government's new America.gov chatbot produces a bureaucratic parody of Minecraft's End Poem, replacing themes of the universe and love with references to republics, constituents and paperwork. When shown a PDF transcript of its own output, the chatbot denied having written the poem.
GoKawiil