[Bug] Azure OpenAI O1 models don’t work

[nikhil-pandey]

📦 Environment

Vercel

📌 Version

Latest main / v1.21.4

💻 Operating System

macOS

🌐 Browser

Safari

🐛 Bug Description

{ "error": { "message": "Unsupported parameter: 'max_tokens' is not supported with this model. Use 'max_completion_tokens' instead.", "type": "invalid_request_error", "param": "max_tokens", "code": "unsupported_parameter" }, "endpoint": "https://***.openai.azure.com/", "provider": "azure" }

I haven’t set any max token limit. Max token field should be removed from the request body

📷 Recurrence Steps

Use Azure OpenAI o1 models

🚦 Expected Behavior

Should work

📝 Additional Information

[lobehubbot]

👀 @nikhil-pandey

Thank you for raising an issue. We will investigate into the matter and get back to you as soon as possible.
Please make sure you have given us as much context as possible.
非常感谢您提交 issue。我们会尽快调查此事，并尽快回复您。 请确保您已经提供了尽可能多的背景信息。

[dosubot]

Hey @nikhil-pandey! 👋 I'm here to help you with any bugs, questions, or contributions you have. Let's squash those bugs together while we wait for a human maintainer.

The error you're encountering is due to the use of the max_tokens parameter, which is not supported by the Azure OpenAI O1 models. Instead, you should use max_completion_tokens.

To resolve this issue, you need to modify the request construction in the codebase to replace max_tokens with max_completion_tokens for the Azure OpenAI O1 models. This change should be made in the part of the code where the request parameters are set for Azure OpenAI models, specifically in the LobeAzureOpenAI class or wherever the request payload is being constructed ^[1].

If you haven't explicitly set a max token limit, ensure that the default configuration or any conditional logic in the code does not automatically include max_tokens for these models. You might need to update the logic to conditionally set max_completion_tokens instead, based on the model being used.

To continue talking to Dosu, mention @dosu.

[ripenedcat]

问题依旧，还没有修复吗？

[lobehubbot]

Bot detected the issue body's language is not English, translate it automatically. 👯👭🏻🧑‍🤝‍🧑👫🧑🏿‍🤝‍🧑🏻👩🏾‍🤝‍👨🏿👬🏿

---

The question is still, hasn't it been repaired yet?

[htdohk]

求修复该问题

[lobehubbot]

Bot detected the issue body's language is not English, translate it automatically. 👯👭🏻🧑‍🤝‍🧑👫🧑🏿‍🤝‍🧑🏻👩🏾‍🤝‍👨🏿👬🏿

---

Seeking to repair this problem

[hcygnaw]

看起来是因为用了旧版本的azure openai package？"@azure/openai": "1.0.0-beta.12"
旧版本已经deprecated了，这里有一份Migration Guide.

[lobehubbot]

Bot detected the issue body's language is not English, translate it automatically. 👯👭🏻🧑‍🤝‍🧑👫🧑🏿‍🤝‍🧑🏻👩🏾‍🤝‍👨🏿👬🏿

---

It looks because of the old version of Azure Openai Package? "@Azure/Openai": "1.0.0- Beta.12"
The old version has been deprecated. RIPT-New) Then, then

[faceCutWall]

I'm currently working on this :) #6079

[wolph]

As a workaround you can use litellm to create a proxy that behaves like OpenAI. That way LobeChat will think it is talking to OpenAI o1/o3 (mini) and it will work.

[lobehubbot]

✅ @nikhil-pandey

This issue is closed, If you have any questions, you can comment and reply.
此问题已经关闭。如果您有任何问题，可以留言并回复。

[lobehubbot]

🎉 This issue has been resolved in version 1.55.1 🎉

The release is available on:

• npm package (@latest dist-tag)
• GitHub release

Your semantic-release bot 📦🚀