Plans for chat completions API / gpt-3.5-turbo?

[ianmacartney]

First of all thank you for providing this in the first place, I'm curious if anyone has started working on the APIs available (like chat completions) with the newer, cheaper gpt-3.5-turbo model? Any forks you know of? I'm less interested in the completion api. I'm guessing it's less useful if it only supports openai, not other models. Thanks, Ian

[r2d4]

Happy to work on this or accept a PR. What is the use case you intend? Small bundle size? What other Chat APIs could llm.ts build an interface over?

[ianmacartney]

My use case is a small bundle that can run in a non-node environment (that doesn't depend on node-fetch, axios, or other libraries). I thought there would be other chat APIs, but they seem to mostly be self-hosted or not expose an API, like hugging chat or mlc-ai/web-llm.