@Nish Are you seeing the same errors for your resource if you are using the speech studio Audio content creation tool?
I think you should be able to save a file of upto 20k characters for S0 tier and 3k characters for F0 tier from the studio that can be synthesized.
I have seen cases where the change from one tier to another i.e F0 to S0 takes time or there is a lag when speech resource auto-scales to support scenarios with large data. If this is an error that is consistent you can raise a support case to review your quota and try again.