AI Product Updates: New Features and Releases

A menu bar tool that switches models for Codex, Claude Code, Gemini CLI and others, with a local gateway translating API protocols and rotating across multiple accounts and keys with failover.

An AI agent generates static, UGC and VSL ads from a single prompt, scans competitor ads, pushes creatives to a Meta ad account and reviews data by the minute without human input.

Google's Antigravity SDK now supports local AI models, letting developers run models locally, with setup details on the official blog.

Built on Google AI Edge's LiteRT, the Antigravity SDK can run open models like Gemma 4 offline on a local GPU, with no API costs and no internet required.

OpenAI says ChatGPT Voice can now use plugins such as email, calendar and Slack, and can be powered by GPT-6, as an official feature update.

OpenAI's voice feature gets a major upgrade: it can now use tools and is available in Work.

A demo shows GPT-Live-1 handling conversation while a Raspberry Pi sends research and tool tasks to GPT-5.6 Luna via the Responses API.

Aspect announced that infinite storage is now free for everyone, six months after launching for teams and enterprises.

Xiaomi's new open-weights flagship model MiMo-V2.6-Pro was added to Wally on its first day and ranks high on the Intelligence Index.

Qwen introduced Planner, Mobile-Use Agent and Creative Agent for smartphones, claiming a 90% end-to-end success rate across apps and an image from one sentence in about 3 seconds.

GPT-Voice now supports plugins for email, calendar and Slack, which the author says makes it far more useful and a real option for daily use.

OpenResearch now supports PubMed, letting users search over 40 million biomedical papers with figures and in-line citations.