[GH-ISSUE #16928] Parameter "think (Ollama)" doesn't seem to work #72700

New Issue

GiteaMirror · 2026-05-13T04:51:36-05:00

GiteaMirror commented

2026-05-13 04:51:36 -05:00

Originally created by @lapo-luchini on GitHub (Aug 26, 2025).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/16928

Check Existing Issues

I have searched the existing issues and discussions.
I am using the latest version of Open WebUI.

Installation Method

Docker

Open WebUI Version

v0.6.25

Ollama Version (if applicable)

0.11.7

Operating System

macOS Darwin 24.5.0

Browser (if applicable)

Firefox 142.0

Confirmation

I have read and followed all instructions in README.md.
I am using the latest version of both Open WebUI and Ollama.
I have included the browser console logs.
I have included the Docker container logs.
I have provided every relevant configuration, setting, and environment variable used in my setup.
I have clearly listed every relevant configuration, custom setting, environment variable, and command-line option that influences my setup (such as Docker Compose overrides, .env values, browser settings, authentication configurations, etc).
I have documented step-by-step reproduction instructions that are precise, sequential, and leave nothing to interpretation. My steps:
Start with the initial platform/version/OS and dependencies used,
Specify exact install/launch/configure commands,
List URLs visited, user input (incl. example values/emails/passwords if needed),
Describe all options and toggles enabled or changed,
Include any files or environmental changes,
Identify the expected and actual result at each stage,
Ensure any reasonably skilled user can follow and hit the same issue.

Expected Behavior

Avoid "thinking".

Actual Behavior

The model did think.

Steps to Reproduce

Open new Qwen3 chat.
Open chat control, click 2 times on "think (Ollama) Default", setting it as "Off".
Write question "test".
See that output starts with "Thought for xx seconds".

Nothing relevant in either the browser console or docker console.

The posted JSON request does seem to have the parameter:

{
  "chat": {
    "models": [
      "qwen3:30b-a3b"
    ],
    "history": {
      // almost the same as "messages"
    },
    "messages": [
      {
        "id": "382edd68-4eed-4f09-ab74-bc7f8a4f0c51",
        "parentId": null,
        "childrenIds": [
          "28ec3404-e7f5-4c21-a5ce-6f8cff8d9e16"
        ],
        "role": "user",
        "content": "test",
        "timestamp": 1756206538,
        "models": [
          "qwen3:30b-a3b"
        ]
      },
      {
        "parentId": "382edd68-4eed-4f09-ab74-bc7f8a4f0c51",
        "id": "28ec3404-e7f5-4c21-a5ce-6f8cff8d9e16",
        "childrenIds": [],
        "role": "assistant",
        "content": "",
        "model": "qwen3:30b-a3b",
        "modelName": "qwen3:30b-a3b",
        "modelIdx": 0,
        "timestamp": 1756206538
      }
    ],
    "params": {
      "think": false // this looks about right?
    },
    "files": []
  }
}

Logs & Screenshots

Additional Information

No response

Originally created by @lapo-luchini on GitHub (Aug 26, 2025). Original GitHub issue: https://github.com/open-webui/open-webui/issues/16928 ### Check Existing Issues - [x] I have searched the existing issues and discussions. - [x] I am using the latest version of Open WebUI. ### Installation Method Docker ### Open WebUI Version v0.6.25 ### Ollama Version (if applicable) 0.11.7 ### Operating System macOS Darwin 24.5.0 ### Browser (if applicable) Firefox 142.0 ### Confirmation - [x] I have read and followed all instructions in `README.md`. - [x] I am using the latest version of **both** Open WebUI and Ollama. - [x] I have included the browser console logs. - [x] I have included the Docker container logs. - [x] I have **provided every relevant configuration, setting, and environment variable used in my setup.** - [x] I have clearly **listed every relevant configuration, custom setting, environment variable, and command-line option that influences my setup** (such as Docker Compose overrides, .env values, browser settings, authentication configurations, etc). - [x] I have documented **step-by-step reproduction instructions that are precise, sequential, and leave nothing to interpretation**. My steps: - Start with the initial platform/version/OS and dependencies used, - Specify exact install/launch/configure commands, - List URLs visited, user input (incl. example values/emails/passwords if needed), - Describe all options and toggles enabled or changed, - Include any files or environmental changes, - Identify the expected and actual result at each stage, - Ensure any reasonably skilled user can follow and hit the same issue. ### Expected Behavior Avoid "thinking". ### Actual Behavior The model did think. ### Steps to Reproduce 1. Open new Qwen3 chat. 2. Open chat control, click 2 times on "think (Ollama) Default", setting it as "Off". 3. Write question "test". 4. See that output starts with "Thought for xx seconds". Nothing relevant in either the browser console or docker console. The posted JSON request does seem to have the parameter: ```json { "chat": { "models": [ "qwen3:30b-a3b" ], "history": { // almost the same as "messages" }, "messages": [ { "id": "382edd68-4eed-4f09-ab74-bc7f8a4f0c51", "parentId": null, "childrenIds": [ "28ec3404-e7f5-4c21-a5ce-6f8cff8d9e16" ], "role": "user", "content": "test", "timestamp": 1756206538, "models": [ "qwen3:30b-a3b" ] }, { "parentId": "382edd68-4eed-4f09-ab74-bc7f8a4f0c51", "id": "28ec3404-e7f5-4c21-a5ce-6f8cff8d9e16", "childrenIds": [], "role": "assistant", "content": "", "model": "qwen3:30b-a3b", "modelName": "qwen3:30b-a3b", "modelIdx": 0, "timestamp": 1756206538 } ], "params": { "think": false // this looks about right? }, "files": [] } } ``` ### Logs & Screenshots <img width="1361" height="1599" alt="Image" src="https://github.com/user-attachments/assets/65844b46-a281-46c2-91a2-1554f4680bd9" /> ### Additional Information _No response_

GiteaMirror added the bug label 2026-05-13 04:51:36 -05:00

GiteaMirror closed this issue

2026-05-13 04:51:37 -05:00

GiteaMirror commented

2026-05-13 04:51:39 -05:00

@tjbck commented on GitHub (Aug 26, 2025):

The model itself also needs to support the parameter (e.g. qwen3:14b).

@tjbck commented on GitHub (Aug 26, 2025): The model itself also needs to support the parameter (e.g. qwen3:14b).

GiteaMirror commented

2026-05-13 04:51:40 -05:00

@lapo-luchini commented on GitHub (Aug 26, 2025):

It is in this list, does it support more than that?

@lapo-luchini commented on GitHub (Aug 26, 2025): It is [in this list](https://ollama.com/search?c=thinking), does it support more than that?

GiteaMirror commented

2026-05-13 04:51:42 -05:00

@rgaricano commented on GitHub (Aug 26, 2025):

It is in this list, does it support more than that?

Also is possible download other gguf models from huggingface or download it manually...

@rgaricano commented on GitHub (Aug 26, 2025): > It is [in this list](https://ollama.com/search?c=thinking), does it support more than that? Also is possible download other gguf models [from huggingface](https://huggingface.co/docs/hub/ollama) or download it manually...

Sign in to join this conversation.

Branches Tags

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: github-starred/open-webui#72700