No one has discussed the test model for a long time. What are your opinions?

The free period of grok-code-fast-1 is ending soon. Do you have any regrets? I think it’s quite suitable for writing one-time scripts, but I never let it handle logical tasks.

As for code-supernova-1-million, this model has an extremely long context. Its publisher hasn’t been announced yet, and does it belong to the Grok family? I used it for a while, but it didn’t leave a deep impression.

Cheetah is the most mysterious model. Surprisingly, it’s a paid-test model. I tried it casually, but its extreme mystery combined with the payment requirement made me feel distrustful, so I stopped using it.