OpenAI o1 is now available in GitHub Copilot Chat in public preview

[Akash1134]

👀 Big news—we’re opening up access to OpenAI’s o1-preview and o1-mini models for everyone in GitHub Copilot. You can now test them out in VS Code, Visual Studio, GitHub.com, and the GitHub Models playground. 🎉

So, what’s the story with the o1-preview and o1-mini models ? Think of these models as your ultra-smart coding sidekicks—they thrive on tackling complex problems, hunting down tricky edge cases, and suggesting thoughtful solutions. Basically, it’s like having an AI buddy that just clicks with your code.

Starting today, you get to play around with both models and switch between them on the fly! Whether you’re breaking down an API, whipping up some code snippets, or untangling a gnarly bug, pick the model that fits the moment and let them bring your ideas to life. 🧩✨

But hey, we want to hear from you—jump in, test them out, and voice your feedback! What’s working, what’s surprising, and where can we make things even better? Let’s build something amazing together. 🚀

Explore how we’re bringing developer choice to copilot with anthropic’s Claude 3.5 sonnet, Google’s Gemini 1.5 pro, and OpenaAI’s O1-preview.

[armesha]

Are there any daily limits for the o1-preview and o1-mini models?

[dstampher]

I haven't been able to find this information anywhere :-(

Hopefully we can find an answer soon.

[GitMurf]

I started getting an error 429 after using o1-preview for not very long... maybe 10 messages today. In googling it, 429 appears to be a rate limiting error so seems like there is some sort of daily limit and unfortunately seems pretty small :-(

[dstampher]

I started getting an error 429 after using o1-preview for not very long... maybe 10 messages today. In googling it, 429 appears to be a rate limiting error so seems like there is some sort of daily limit and unfortunately seems pretty small :-(

The current limit for ChatGPT Plus users is 50 per week for o1-preview, and 50 per day for o1-mini.

So, even if it is 10 per day with o1-preview, that would still be significantly more per month than even a paid ChatGPT sub.

EDIT: After some digging, I seem to have finally found some usage limits, but they are for the Github Models playground. However, based on the fact that you said you got 429 errors after ~10 o1-preview messages, it seems that these usage limits might also apply to Copilot: https://docs.github.com/en/github-models/prototyping-with-ai-models#rate-limits

In which case the limit would be 8 per day with o1-preview, and 12 per day with o1-mini.

[sheldonhull]

Does this require accepting preview of models?
My org won't accept preview/early access, so not sure if I should be able to still use o1-preview without opting into the preview of models for this.
I have azure openai o1-preview as well, but I haven't seen any mention I can track down of being able to connect my own enterprise deployment to it.

[samus-aran]

Hi @sheldonhull 👋

For both o1s your organisation would need to accept the preview access. If you have copilot access via your enterprise/org and the enterprise/org has disabled it, you would not be able to use it.

[kangkannnng]

Can we use api to get access?

[samus-aran]

Hi @kangkannnng, unfortunately not. API usage of these are only currently available with our integration partners. Unauthorised use will incur to warnings and suspensions.
Alternatively you can use the API via agents! Here are the docs: https://docs.github.com/en/copilot/building-copilot-extensions/building-a-copilot-agent-for-your-copilot-extension/using-copilots-llm-for-your-agent. Additionally, you are able to experiment using GitHub Models which does have o1 available: https://github.com/marketplace/models 👍

[morty6688]

When can we use o1 in jetbrains IDE?