I’ve been working on a collection of chat persona’s the last several weeks now that use GPT-3.5-Turbo, and today I saw GPT-3.5-Turbo-1106 available in the model list. I plugged this into my API, and my god, the behavior and responses from the GPT bots are a million times better. I was just expecting very slight improvements to the responses here and there, but it has far surpassed my expectations. It has truly made my day.
On a side note though, it seems more common now that I send a prompt and get no response. I won’t complain too much. I can probably come up with an automated retry system if I fail to receive a response in x amount of time. It’s just a waste of tokens, but at $0.001 per 1k tokens, this is acceptable.