According to Vetted Consumer, a new open-weight AI model called GLM-5.2 just became the highest-ranked model on the Artificial Analysis Intelligence Index. Built by Chinese lab Z.ai, it's capable of handling a million-token context window — roughly equivalent to a small book in a single prompt. The catch? The full model weighs one point five one terabytes. Running it locally is... ambitious. You'd realistically need a Mac Studio M3 Ultra with two hundred fifty-six gigabytes of unified memory — a nine thousand five hundred dollar machine that generates about three to nine tokens per second. For most people, using the API or cloud rental makes far more sense. The model is built for serious agentic coding and long-context work, where its efficiency innovations really shine.

Listen to this story

Hear this and more stories in a personalized audio briefing.

Open The Chonkerton