Z.ai has released some new models that are very impressive on benchmarks https://z.ai/blog/glm-4.5 They also have very low costs, claiming to cost quite a bit less than Deepseek models while besting their performance.

API details: https://z.ai/model-api

* GLM-4.5

* GLM-4.5-Air

* GLM-4.5-Flash

(For other reference points: https://openrouter.ai/z-ai/glm-4.5, https://huggingface.co/zai-org/GLM-4.5)

Given some of the recent friction with fair limits, adding these models should give a very cost effective way of getting super level performance at a very reasonable cost. I'd suggest you integrate these as soon as possible.