Anecdotally, as a roleplaying chat experience, char.ai seems to perform way better than anything else publicly available (doesn't get repetitive, very long memory). It also feels different to GPT3 on how it is affected by prompts.
I've just assumed that char.ai is doing its own thing as it was founded by two engineers who worked on google's LaMDA.
