Open Weight vs Low Cost. Are these models cheap because they are open weight and having hundreds or people stress test running them on different hardware helped to lower the cost? Or is it that they are being provided as loss leaders to drive the prices down?

I wouldn’t be surprised if the A.I. in China is heavily subsided so that people use them. Not saying that’s necessarily a bad thing, but I’m not sure it’s sustainable long term, or a fair way to compare A.I. models. We need to compare the true computational and financial cost of each token internally, not whatever is being charged for it externally, otherwise you’re just saying “stuff made in China is generally cheaper” which we’ve known for decades. This might be a complicated question to answer though.