GLM 5.2 Challenges Closed AI Models | Durham News

China’s Zhipu AI just dropped GLM 5.2 — an open-source powerhouse that’s now matching Anthropic’s top-tier Opus 4.8 on benchmarks, but for a fraction of the cost. While Zhipu claims its model can rival the secretive Claude Mythos in spotting security flaws, that’s hard to verify since Mythos is off-limits to most. Meanwhile, GLM 5.2’s popularity is exploding, with developers embracing its ability to handle complex tasks like coding and testing — all while giving companies control over their AI infrastructure. With rising AI costs and government restrictions tightening around closed models, open models like GLM 5.2 are becoming the smart, budget-friendly choice — especially as access to elite models like Mythos is now limited to just 100 vetted organizations. Listen in comfort:Get a discount on a Soli Pillow: http://solipillow.com/discount/dnn Advertise on DNN:[email protected] This is an automated, high-level news summary based on public reporting.Report issues to [email protected]. View sources & latest updates:https://sources.thednn.ai/8d0b13950d3b6275