microsoft
GPT-4o in Copilot
We'll update this issue as we get access to 4o, meanwhile please subscribe and add reactions.
To keep this issue focused on model update: If you got feature requests inspired by GPT-4o's new capabilities, please file a new issue.
Also there is this topic here:
https://github.com/orgs/community/discussions/123670
I tried to ask copilot in vscode which GPT version it uses. I've got this answer:
I am using the GPT-4o version of OpenAI's GPT models.
So seems like copilot is already using GPT-4o model in vscode. EDIT: I just checked, and it's officially in the release notes for the VSCode update 1.92 that Copilot uses GPT-4o. To use this model, you need to upgrade to the latest version, 1.92.
Copilot uses GPT-4o - GitHub Copilot Chat upgraded to OpenAI's GPT-4o.
I tried to ask copilot in vscode which GPT version it uses. I've got this answer:
I am using the GPT-4o version of OpenAI's GPT models.So seems like copilot is already using GPT-4o model in vscode. EDIT: I just checked, and it's officially in the release notes for the VSCode update 1.92 that Copilot uses GPT-4o. To use this model, you need to upgrade to the latest version, 1.92.
Copilot uses GPT-4o - GitHub Copilot Chat upgraded to OpenAI's GPT-4o.
Copilot Chat has been upgraded to 4o model.
What about GitHub Copilot? Which model is it using? I'm referring to the Copilot autocomplete, not chat.
What about GitHub Copilot? Which model does it use? I'm talking about the Copilot autocomplete feature, not the chat function.
It's still 3.5, but it seems like Microsoft is experimenting with the 4o mini, so we might expect some upgrades in the future.
GitHub Copilot Chat (v0.17+) has been updated to utilize GPT-4o, given you much faster responses with higher quality code: https://aka.ms/github-copilot-chat-gpt-4o
GitHub Copilot Chat (v0.17+) has been updated to utilize GPT-4o, given you much faster responses with higher quality code: https://aka.ms/github-copilot-chat-gpt-4o
I've tried GPT-4o-mini for auto complete and it works fine, but in one day of basic coding it used 1.5 Million tokens and 1500 api calls. Daily cost about 25 cents. Maybe is still cheaper than the github copilot. The results were ok and fast.
GitHub Copilot Chat (v0.17+) has been updated to utilize GPT-4o, given you much faster responses with higher quality code: https://aka.ms/github-copilot-chat-gpt-4o
I've tried GPT-4o-mini for auto complete and it works fine, but in one day of basic coding it used 1.5 Million tokens and 1500 api calls. Daily cost about 25 cents. Maybe is still cheaper than the github copilot. The results were ok and fast.
@mcpdigital How have you managed to use GPT-4o-mini for autocomplete without GitHub Copilot?
