What I find remarkable is that deepseek and qwen are much more open about the model output (not hiding intermediate thinking process), open their weights, and a

What I find remarkable is that deepseek and qwen are much more open about the model output (not hiding intermediate thinking process), open their weights, and a lot of time, details on how they are trained, and the caveats along the way. And they don't have "Open" in their names.