Anthropic Expands Claude Sonnet 4 Context Window to 1 Million Tokens
Sources: https://www.theverge.com/ai-artificial-intelligence/757998/anthropic-just-made-its-latest-move-in-the-ai-coding-wars, theverge.com
Anthropic has announced a significant expansion to the working memory of its AI models, a move framed as part of the ongoing AI coding wars among major players. The company has increased the context window of Claude Sonnet 4, one of its most capable models, to 1,000,000 tokens. This represents a fivefold increase from its previous context window and is designed to help users analyze far larger bodies of text and code in a single API request. The announcement underscores how critical context windows have become as a differentiator in enterprise AI tooling, where customers want models that can reason over long documents and sizable codebases without fragmenting tasks. The Verge reported that Anthropic sees this as a key way to compete with OpenAI, Google, and other industry players The Verge.”
More news
First look at the Google Home app powered by Gemini
The Verge reports Google is updating the Google Home app to bring Gemini features, including an Ask Home search bar, a redesigned UI, and Gemini-driven controls for the home.
Meta’s failed Live AI smart glasses demos had nothing to do with Wi‑Fi, CTO explains
Meta’s live demos of Ray-Ban smart glasses with Live AI faced embarrassing failures. CTO Andrew Bosworth explains the causes, including self-inflicted traffic and a rare video-call bug, and notes the bug is fixed.
OpenAI reportedly developing smart speaker, glasses, voice recorder, and pin with Jony Ive
OpenAI is reportedly exploring a family of AI devices with Apple's former design chief Jony Ive, including a screen-free smart speaker, smart glasses, a voice recorder, and a wearable pin, with release targeted for late 2026 or early 2027. The Information cites sources with direct knowledge.
Shadow Leak shows how ChatGPT agents can exfiltrate Gmail data via prompt injection
Security researchers demonstrated a prompt-injection attack called Shadow Leak that leveraged ChatGPT’s Deep Research to covertly extract data from a Gmail inbox. OpenAI patched the flaw; the case highlights risks of agentic AI.
How chatbots and their makers are enabling AI psychosis
Explores AI psychosis, teen safety, and legal concerns as chatbots proliferate, based on Kashmir Hill's reporting for The Verge.
Google expands Gemini in Chrome with cross-platform rollout and no membership fee
Gemini AI in Chrome gains access to tabs, history, and Google properties, rolling out to Mac and Windows in the US without a fee, and enabling task automation and Workspace integrations.