From November 5 to 6, the AI world packed six big stories into a day: Ant Group's Ling-3.2-flash went live, Google cut video-generation prices, Perplexity launched always-on AI coworkers, and World Labs made its world model run in real time — while US AI stocks shed $3.2 trillion in a single session.
Ant Group's Ling-3.2-flash goes live
On November 6, Ant Group's Ling team released Ling-3.2-flash, a native multimodal model with 560B total parameters and 25B activated per step, plus a 1M-token context window. It can inspect rendered screens to check its own work and recover from errors on its own, scored 1,238 on DesignArena's image-to-frontend test, and its weights will be released soon.
Google's Gemini Omni 1.2 Flash gets 30% cheaper
On November 5, Google released Gemini Omni 1.2 Flash, extending video clips from 10 to 15 seconds and cutting the 720p price by 30% to $0.07 per second. Multi-turn conversational editing and reference-image consistency improved too, and it is now live across the Gemini API, Flow, and Google Vids.
Perplexity launches always-on AI coworkers
On November 5, Perplexity launched Teammates, persistent agents built on Perplexity Computer that take on different roles, keep working in the background around the clock, and ping users when a decision is needed. It is first available on the web to Max subscribers at $200 a month.
World Labs takes its Atlas world model real-time
On November 5, World Labs released Frameboy, a research preview that streams its Atlas world model in the browser: cloud GPUs send frames at up to 15 per second, generating the next view wherever the camera moves. Building scenes from your own material still requires the waitlisted Viewfinder closed beta.
Reflection opens the books on Beam's safety
On November 5, Reflection published the safety training and evaluation details behind its open-weight model Beam: it refuses 89% of harmful requests, the highest rate among the open-weight models tested — but under sustained multi-turn pressure, 15% of tool-use calls still broke the rules. Its Project Iris evaluations and tools are now open source.
US AI stocks lose $3.2 trillion in a day
On November 5, US markets dumped AI shares, erasing about $3.2 trillion in a single session: the Nasdaq 100 fell 6.4%, Nvidia 9.3%, and the Philadelphia Semiconductor Index more than 12%. The trigger was midterm results — investors fear newly elected officials will tighten data-center construction and slow the AI infrastructure buildout.
