-
Notifications
You must be signed in to change notification settings - Fork 760
wip: migrate to any llm #6024
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
base: main
Are you sure you want to change the base?
wip: migrate to any llm #6024
Conversation
|
The latest updates on your projects. Learn more about Vercel for Git ↗︎
|
|
Hi! I'm on the any-llm team. Feel free to let me know if there are features not yet supported in any-llm that you need. I see the ticket mentions lack of "thinking" support, not sure if you're referencing thinking as an input parameter or as an output item. |
|
hey @njbrake, it did not seem like any reasoning tokens were returned. Could be an issue in our code, but I couldn't get reasoning/thinking tokens from the completions API |
|
Ah, I figured it out: we don't support reasoning tokens in all providers yet: see https://mozilla-ai.github.io/any-llm/providers/ for the compatibility matrix. OpenAI (the provider) doesn't return reasoning tokens in their completion api, only in their responses API. See https://blog.mozilla.ai/standardized-reasoning-content-a-first-look-at-using-openais-gpt-oss-on-multiple-providers-using-any-llm/ for the explanation of those differences. For anthropic and google, we don't support it yet but I believe those providers do return reasoning information, so we would only need to implement the conversion logic to be able to return it. |
|
@njbrake awesome, thank you! will definitely upgrade a try it out. |
This removes fragile logic for converting to different LLMs and instead rely on https://github.com/mozilla-ai/any-llm.
Some oustanding blocking issues: