China’s Z.ai claims it can match Mythos on cybersecurity
China's Zhipu AI (Z.ai) released its open-weight GLM-5.2, and some researchers have claimed that it matches Mythos in certain bug-finding and cybersecurity scenarios. While GLM lags behind models from Anthropic and OpenAI in other, more general tasks, it seems that China has dramatically reduced the gap in the capabilities between its models and those of the US. This level of advancement is particularly concerning to the US government, which has worked to restrict China's acces
Get smart on it
China's Zhipu AI released an open-weight model that some researchers claim matches a US competitor's performance on bug-finding and cybersecurity tasks, though it still lags on general tasks. The US government has restricted China's access to advanced AI models and training hardware due to national security concerns, viewing such models as threats because of their ability to identify vulnerabilities. Because this Chinese model is open-weight, anyone can download and run it on standard hardware, which provides flexibility for legitimate users but also creates risks for misuse by bad actors operating with little oversight.
Related news
DeepReinforce Open Sources Ornith-1.0 Coding Models - Open Source For You
DeepReinforce has open sourced its Ornith-1.0 coding model family, releasing model weights and research on Hugging Face to help developers build, study and
Liquid AI Ships LFM2.5-230M with llama.cpp, MLX, vLLM, SGLang, and ONNX Support for On-Device Inference
Liquid AI shipped LFM2.5-230M, it’s the company’s smallest model to date. The release targets a specific job: running agentic tasks on phones, robots, and automation devices. Both the base and instruction-tuned checkpoints are open-weight on Hugging Face. The pitch is narrow on purpose. This is not a general reasoning model. It is built for data extraction and tool use on edge hardware. TL;DR Liquid AI’s LFM2.5-230M is its smallest model yet: 230M params, open-weight
