404 Error: GitHub AI Inference API endpoint not found for Cohere embedding model

[knamit]

Select Topic Area

Bug

Body

Issue Description

When attempting to use the GitHub AI Inference API with the Cohere embedding model, I'm receiving a 404 "page not found" error. The issue appears to be that the endpoint URL is auto-redirecting to /v1/embed when it should redirect to /v3/embed for the Cohere v3 model.

Expected Behavior

The API should automatically redirect to the correct version endpoint (/v3/embed) for the Cohere v3 embedding model, or properly handle the request at the v1 endpoint.

Actual Behavior

The API auto-redirects to /v1/embed instead of /v3/embed, resulting in a 404 error with the message "404 page not found".

API Details

• Endpoint: https://models.github.ai/inference
• Model: cohere/Cohere-embed-v3-english
• Incorrect auto-redirect: /v1/embed
• Expected redirect: /v3/embed
• Full URL that fails: https://models.github.ai/inference/v1/embed

Error Response

{
  "ok": false,
  "error": {
    "reason": "non-json",
    "statusCode": 404,
    "rawBody": "404 page not found\n"
  }
}

Root Cause

The endpoint routing logic appears to be incorrectly auto-redirecting all embedding requests to the v1 API endpoint, regardless of the model version specified. For Cohere v3 models, requests should be routed to the v3 endpoint.

Questions

Is the auto-redirect behavior intentional, and should v1 endpoints support v3 models?
Is there a way to explicitly specify the API version in the request to bypass the auto-redirect?
Are there updated documentation or examples for using Cohere v3 models with the GitHub AI Inference service?

[github-actions[bot]]

💬 Your Product Feedback Has Been Submitted 🎉

Thank you for taking the time to share your insights with us! Your feedback is invaluable as we build a better GitHub experience for all our users.

Here's what you can expect moving forward ⏩

• Your input will be carefully reviewed and cataloged by members of our product teams. 
  • Due to the high volume of submissions, we may not always be able to provide individual responses.
  • Rest assured, your feedback will help chart our course for product improvements.
• Other users may engage with your post, sharing their own perspectives or experiences. 
• GitHub staff may reach out for further clarification or insight. 
  • We may 'Answer' your discussion if there is a current solution, workaround, or roadmap/changelog post related to the feedback.

Where to look to see what's shipping 👀

• Read the Changelog for real-time updates on the latest GitHub features, enhancements, and calls for feedback.
• Explore our Product Roadmap, which details upcoming major releases and initiatives.

What you can do in the meantime 💻

• Upvote and comment on other user feedback Discussions that resonate with you.
• Add more information at any point! Useful details include: use cases, relevant labels, desired outcomes, and any accompanying screenshots.

As a member of the GitHub community, your participation is essential. While we can't promise that every suggestion will be implemented, we want to emphasize that your feedback is instrumental in guiding our decisions and priorities.

Thank you once again for your contribution to making GitHub even better! We're grateful for your ongoing support and collaboration in shaping the future of our platform. ⭐

[LindseyB]

This is definitely a bug. Thanks for the report it's been passed on to the team.

[garman]

Thank you for reporting this!

Could you provide more information about how you're using this endpoint? Are you using an SDK or crafting these URL's yourself? If you have control over the URL you're using in the request try changing it to https://models.github.ai/inference/embeddings -- Note the lack of a version and the longer-form embeddings. Our REST API docs for embedding models/requests has some more information too!

If using an SDK, which one?

[knamit]

I am using the Cohere AI SDK the endpoint that I was using was https://models.github.ai/inference

[garman]

Thank you!

I'm able to recreate the error using the Cohere AI SDK. A workaround for the time being would be to use the Azure AI Inference SDK for python

An example that I was able to confirm worked:

pip install azure-ai-inference

import os

from azure.ai.inference import EmbeddingsClient
from azure.core.credentials import AzureKeyCredential

endpoint = "https://models.github.ai/inference"
model_name = "cohere/Cohere-embed-v3-english"
token = os.environ["GITHUB_TOKEN"]

client = EmbeddingsClient(
    endpoint=endpoint,
    credential=AzureKeyCredential(token)
)

response = client.embed(
    input=["first phrase", "second phrase", "third phrase"],
    model=model_name
)

for item in response.data:
    length = len(item.embedding)
    print(
        f"data[{item.index}]: length={length}, "
        f"[{item.embedding[0]}, {item.embedding[1]}, "
        f"..., {item.embedding[length-2]}, {item.embedding[length-1]}]"
    )
print(response.usage)