Chinese AI Ships Open Source LLM, Beats Competitors

Chinese AI startup Z.ai has released GLM-5.1, an open-source large language model capable of autonomous operation for up to eight hours on a single task. This advanced Mixture-of-Experts model, boasting 754 billion parameters, significantly outperforms competitors like GPT-5.4 and Claude Opus 4.6 on benchmarks such as SWE-Bench Pro. GLM-5.1 utilizes a novel 'staircase pattern' optimization, allowing it to iteratively refine solutions and overcome complex engineering challenges autonomously. The model is available under a permissive MIT License, fostering wider adoption and commercial use within the developer community.