A comparison of open large language models is underway, with various outlets discussing the strengths of different models. The models being discussed include Llama 4, Qwen 3.5, and Mistral, as well as others such as Gemma 4 and Claude Sonnet 5. The conversation also involves the development and optimization of these models, including efforts by Chinese labs and companies like NVIDIA and Google.