GLM-5.2 Beats GPT-5.5 Pro at 1/6th Cost: New AI Model

A Chinese AI company, Z.ai, has released a new open-weight model called GLM-5.2. This model, with 753 billion parameters, is designed for autonomous coding and engineering tasks. Here's the thing: GLM-5.2 scored 62.1 on SWE-bench Pro, surpassing GPT-5.5, which scored 58.6. What's interesting is that it achieved this at a reported one-sixth of the cost of GPT-5.5 Pro. The model supports a stable one-million-token context window and is available through various platforms. Z.ai has released the model weights under an MIT license, allowing companies to download, modify, and deploy it on their own infrastructure. The model introduces an architecture called IndexShare, which helps reduce computing requirements. It also offers "Max" and "High" thinking modes, letting users prioritize performance or reduce token usage. On long engineering tasks, GLM-5.2 showed strong performance, including 34.3% on PostTrainBench, compared to 28.4% for GPT-5.5. The bottom line is that this open-source model could offer a powerful and cost-effective alternative for businesses and developers.