[GH-ISSUE #3022] [BUG]: Switching model for workspace not working #1935

Closed
opened 2026-02-22 18:27:18 -05:00 by yindo · 1 comment
Owner

Originally created by @marquies on GitHub (Jan 24, 2025).
Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/3022

How are you running AnythingLLM?

AnythingLLM desktop app

What happened?

I want to use Anything LLM with ollama for a specific Workspace. But it looks like it falls back to OpenAI Chat GPT.

Steps to reproduce:
0. Setup ChatGPT Key

  1. Open Anything LLM
  2. Create new Workspace
  3. Open Workspace settings, select Chat-Settings Tab, select OLLAMA and Model
  4. Open new Chat Threat in the project, Prompt "Which model are you?"

Expected Result: for example something like "I'm Qwen 2.5 model"
Actual Result: something like "I am based on OpenAI’s GPT-4 model. How can I assist you today?"

Version 1.7.2 Mac Silicon.

Image

Are there known steps to reproduce?

No response

Originally created by @marquies on GitHub (Jan 24, 2025). Original GitHub issue: https://github.com/Mintplex-Labs/anything-llm/issues/3022 ### How are you running AnythingLLM? AnythingLLM desktop app ### What happened? I want to use Anything LLM with ollama for a specific Workspace. But it looks like it falls back to OpenAI Chat GPT. Steps to reproduce: 0. Setup ChatGPT Key 1. Open Anything LLM 2. Create new Workspace 3. Open Workspace settings, select Chat-Settings Tab, select OLLAMA and Model 4. Open new Chat Threat in the project, Prompt "Which model are you?" Expected Result: for example something like "I'm Qwen 2.5 model" Actual Result: something like "I am based on OpenAI’s GPT-4 model. How can I assist you today?" Version 1.7.2 Mac Silicon. ![Image](https://github.com/user-attachments/assets/703e5008-b474-4893-9fe7-79229bca3bd5) ### Are there known steps to reproduce? _No response_
yindo added the needs info / can't replicateDesktop labels 2026-02-22 18:27:18 -05:00
yindo closed this issue 2026-02-22 18:27:18 -05:00
Author
Owner

@timothycarambat commented on GitHub (Jan 24, 2025):

Expected Result: for example something like "I'm Qwen 2.5 model"
Actual Result: something like "I am based on OpenAI’s GPT-4 model. How can I assist you today?"

This is a really faulty way of testing what model is being used for inference. Models do not know who they are and they very well could answer randomly like this even if the model is something else.

Open Workspace settings, select Chat-Settings Tab, select OLLAMA and Model

Did you scroll to the bottom and click save to perist that change? The button is pushed to the bottom (bad UX) - I presume you saved but it wasn't explicitly mentioned.

Outside of this, cannot replicate.

@timothycarambat commented on GitHub (Jan 24, 2025): > Expected Result: for example something like "I'm Qwen 2.5 model" > Actual Result: something like "I am based on OpenAI’s GPT-4 model. How can I assist you today?" This is a really faulty way of testing what model is being used for inference. Models do not know who they are and they very well could answer randomly like this even if the model is something else. > Open Workspace settings, select Chat-Settings Tab, select OLLAMA and Model Did you scroll to the bottom and click save to perist that change? The button is pushed to the bottom (bad UX) - I presume you saved but it wasn't explicitly mentioned. Outside of this, cannot replicate.
yindo changed title from [BUG]: Switching model for workspace not working to [GH-ISSUE #3022] [BUG]: Switching model for workspace not working 2026-06-05 14:43:32 -04:00
Sign in to join this conversation.
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: Mintplex-Labs/anything-llm#1935