Perplexity released a research preview of a post-trained version of Z.AI’s GLM 5.2, built to operate inside its Computer agent harness and available now in production. The system costs one-third the price of Opus 4.8 across benchmarks, delivering near-frontier performance at 0.344x of the cost of Opus. GLM 5.2 is a roughly 744-billion-parameter model from Z.ai—formerly Zhipu AI, a Beijing lab that has been on the U.S. Entity List since January 2025. Released under an MIT license in June, it sits among the top AI models on long-horizon coding benchmarks at a fraction of the API cost. Perplexity used post-training to teach GLM 5.2 a critical skill: knowing when to handle a task itself and when to escalate to something more powerful. The fine-tuned GLM 5.2 includes an "advisor tool"—a native capability to recognize when a query exceeds its own competence and hand off to a third-party frontier model. CEO Aravind Srinivas wrote on X, "When paired with an advisor, this model functions at Opus 4.8 grade performance at a fraction of the cost." It is Perplexity's second Chinese open-source fine-tune in 18 months—the first being R1-1776, a version of DeepSeek R1 stripped of roughly 300 Beijing-mand