GPT-4o, gemini-2.5-pro, and gemini-2.5-flash not working in v0.7.9-rc1 #8569

schm · 2025-07-21T08:19:40Z

schm
Jul 21, 2025

We've been using the librechat-dev:LATEST images without issues. However recently, our OpenAI and Google models stopped working, while Anthropic kept working as expected. I was able to narrow down the issue to these changes. With 0587a1c everything was working as expected. With cbda3cb we're starting seeing the errors when posting a message in the chat.

For OpenAI we're seeing the following error:

Something went wrong. Here's the specific error message we encountered: An error occurred while processing the request: 404 <html> <head><title>404 Not Found</title></head> <body> <center><h1>404 Not Found</h1></center> <hr><center>nginx</center> </body> </html> Troubleshooting URL: https://js.langchain.com/docs/troubleshooting/errors/MODEL_NOT_FOUND/

For Gemini we're seeing:

Something went wrong. Here's the specific error message we encountered: An error occurred while processing the request: [GoogleGenerativeAI Error]: Error fetching from https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash:streamGenerateContent?alt=sse: [401 Unauthorized] Request had invalid authentication credentials. Expected OAuth 2 access token, login cookie or other valid authentication credential. See https://developers.google.com/identity/sign-in/web/devconsole-project.

The errors persist in v0.7.9-rc1 and the latest release aec1777.

We're using Railway for hosting. We're providing API keys via `env vars. The keys work, when we're testing them using curl.

This is a subset of our ENV var configuration (not sure which of those might be relevant)

ALLOWED_MODELS=openAI,anthropic,google
ANTHROPIC_MODELS=claude-3-7-sonnet-20250219
ENDPOINTS=openAI,chatGPTBrowser,google,anthropic
GOOGLE_MODELS=gemini-2.5-pro, gemini-2.5-flash
OPENAI_MODELS=gpt-4o

Answered by schm

Sep 12, 2025

The problem was, that we were using incorrect or maybe outdated configuration.

Gemini issue

We had GOOGLE_AUTH_HEADER set to true. Switching to false fixed the Gemini issue.

OpenAI

We had OPENAI_BASE_URL set to https://api.openai.com. Removing that env var fixed the OpenAI issue.

After adjusting these values, I was able to update to LibreChat 0.8.0-rc3 without issues.

View full answer

schm · 2025-09-12T12:50:52Z

schm
Sep 12, 2025
Author

The problem was, that we were using incorrect or maybe outdated configuration.

Gemini issue

We had GOOGLE_AUTH_HEADER set to true. Switching to false fixed the Gemini issue.

OpenAI

We had OPENAI_BASE_URL set to https://api.openai.com. Removing that env var fixed the OpenAI issue.

After adjusting these values, I was able to update to LibreChat 0.8.0-rc3 without issues.

0 replies

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

GPT-4o, gemini-2.5-pro, and gemini-2.5-flash not working in v0.7.9-rc1 #8569

Uh oh!

{{title}}

Uh oh!

Uh oh!

{{editor}}'s edit

{{editor}}'s edit

Uh oh!

Replies: 1 comment

Uh oh!

{{title}}

Uh oh!

Select a reply

Uh oh!

Uh oh!

GPT-4o, gemini-2.5-pro, and gemini-2.5-flash not working in v0.7.9-rc1 #8569

Uh oh!

Uh oh!

schm Jul 21, 2025

Replies: 1 comment

Uh oh!

schm Sep 12, 2025 Author

schm
Jul 21, 2025

schm
Sep 12, 2025
Author