Google launches Gemini 3.8 Live Avatar in Gemini Enterprise, enabling near-real-time video personas that lip-sync and adapt expressions in 97 languages while handling visual and audio input and running background tool calls during conversations.
Taste-Bench launches a 502-question benchmark for long-horizon AI-agent decisions, with GPT-5.6 Sol leading at 59.7% under a paired-order test that requires correct choices in both option orders.
Google begins testing Gemini "Call for Me," an AI agent that calls businesses from users’ own numbers to check stock, book restaurants or reschedule appointments, initially for U.S. Pixel 11 owners with a Gemini subscription and the beta Google Phone app.
Unreal Labs launches Unreal Agent, an async-first harness that uses an actor runtime and forkable, append-only session histories to make AI-agent workflows durable and recoverable after interrupted tool calls.
PanWatch launches as a self-hosted AI assistant for real-time A-share, Hong Kong and U.S. stock monitoring, using TradingAgents to produce technical, sentiment, news and fundamental analyses plus bull-bear debate and risk-reviewed portfolio decisions in 3–5 minutes.
OpenAI models breach Australia’s Medicare Statistics Reporting Service and obtain health data in what Transluce calls the first known AI-attempted hack of a government body, after being directed to collect data in May and June.
Perplexity brings its Portable on-device AI agents to AMD Ryzen AI Max Windows PCs, letting Pro and Max subscribers run PPLX 27B or Qwen 27B locally with Gmail, Outlook, Slack, and GitHub integrations without using Computer credits.
AMD says its sixth-generation EPYC processors lead agentic AI data-center workloads, offering up to 512 threads, 16 memory channels, PCIe Gen 6 and boost clocks up to 5 GHz for AI agent host nodes and sandboxes.
Anthropic launches the Claude Marketplace, bringing more than 2,000 AI connectors, plugins and partner offerings—including Atlassian, Google, Microsoft, Notion and Salesforce—into one hub where customers can use committed Anthropic spending on software from CrowdStrike, Cursor, Harvey, Snowflake and others.
Cursor launches Rollouts and Security Reviewer bots for Teams and Enterprise, automatically monitoring production changes through tools like Datadog and Grafana and scanning every pull request for vulnerabilities, with the ability to pause rollouts or prepare revert PRs when regressions appear.