Change default reasoning benchmark from BRIGHT(v1) to BRIGHT(v1.1)

### Description of the feature

BRIGHT v1.1's implementation is much cleaner and should be the default - we will have to rerun the models using the new implementation though before we can switch over

Discussed here: https://github.com/embeddings-benchmark/mteb/issues/1517#issuecomment-4148235399