A video model now labels footage shot from a robot's own view.
TwelveLabs released Pegasus 1.6 on Tuesday, its first model built for first-person video. Its chief executive says it processed seventeen years of footage in eighteen hours. The company charges $1.75 per hour of video and published no benchmark scores.
Why it mattersRobots learn chores from video of people working, and labelling it by hand is slow. How fast that labelling gets done sets how soon those robots arrive.
Read in the app — free, in 9 languages
Related stories
Google's new search model fits in 191 megabytes of phone memory.2026-10-07Mistral will give away its trillion-parameter model this month.
2026-10-06Reflection will hand out a 501-billion-parameter model for free.
2026-10-06Microsoft's new transcriber starts writing before you finish speaking.
2026-10-05An American lab is about to give away a China-class open model.
2026-10-05