As new models come out, it seems they're going to discontinue older models and allocate the resources from them to the new ones.
glm5.2 is good, but it produces too many tokens, so I'm mainly using 5.1.
minimax 3 is also decent, but I'm currently using 2.7 for simple coding.
There were too many qwen models.
The usage amount in the subscription plan remains the same, but even though they say the performance of cloud models has improved, the number of tokens used when doing the same task keeps increasing. It's scary ㅠㅠ
