The Chinese startup Zhipu AI has released its flagship language model, GLM-5.2, designed for long-agent tasks and programming. This open-source solution features a context window of 1 million tokens, an MIT license, and supports local deployment.
On its Hugging Face page, the model is described as capable of generating text in both English and Chinese, with a size of 753 billion parameters.
GLM-5.2 offers multiple levels of "reasoning intensity" to balance quality and latency. Its architecture also incorporates IndexShare and an updated MTP layer for speculative decoding.
According to the developers, IndexShare reuses a single indexer for every four layers of sparse attention, reducing the number of operations per token by 2.9 times. The MTP update increases confirmation length by up to 20%.
In three key benchmarks—FrontierSWE, PostTrainBench, and SWE-Marathon—GLM-5.2 outperformed other open-source models.
Source: Zhipu AI.In standard programming performance tests, GLM-5.2 also emerged as the most powerful open-source model.
Source: Zhipu AI.GLM-5.2 is distributed under the open MIT license. It supports local deployment with SGLang, vLLM, Transformers, KTransformers, and Docker Model Runner. Quantization options are available for llama.cpp, Ollama, and LM Studio.
As a reminder, in June, the Rio de Janeiro IT company IplanRIO introduced Rio 3.5 Open 397B as an open AI model funded by public resources. However, a day later, the Nex team claimed that the tool appeared to be a direct merger of Nex-N2-Pro and Qwen3.5-397B-A17B.