I have searched for all existing open AND closed issues and discussions for similar requests. I have found none that is comparable to my request.
Verify Feature Scope
I have read through and understood the scope definition for feature requests in the Issues section. I believe my feature request meets the definition and belongs in the Issues section instead of the Discussions.
Problem Description
with llama.cpp it's possible to disable thinking through the api using : thinking_budget_tokens: 0, there may also be other methods that work with it.
anyway, currently there is no convenient builtin way to disable thinking for llama.cpp
it can be done per chat by adding "thinking_budget_tokens: 0" as a custom field in the chat settings, but there is no way to do it globaly.
Desired Solution you'd like
either add an option for llama.cpp in settings or allow custom fields not only in the chat settings but globaly.
Alternatives Considered
No response
Additional Context
No response
Originally created by @alkeryn on GitHub (May 4, 2026).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/24341
### Check Existing Issues
- [x] I have searched for all existing **open AND closed** issues and discussions for similar requests. I have found none that is comparable to my request.
### Verify Feature Scope
- [x] I have read through and understood the scope definition for feature requests in the Issues section. I believe my feature request meets the definition and belongs in the Issues section instead of the Discussions.
### Problem Description
with llama.cpp it's possible to disable thinking through the api using : thinking_budget_tokens: 0, there may also be other methods that work with it.
anyway, currently there is no convenient builtin way to disable thinking for llama.cpp
it can be done per chat by adding "thinking_budget_tokens: 0" as a custom field in the chat settings, but there is no way to do it globaly.
### Desired Solution you'd like
either add an option for llama.cpp in settings or allow custom fields not only in the chat settings but globaly.
### Alternatives Considered
_No response_
### Additional Context
_No response_
@owui-terminator[bot] commented on GitHub (May 4, 2026):
🔍Similar Issues Found
I found some existing issues that might be related. Please check if any of these are duplicates or contain helpful solutions:
#23703issue: Notes feature not compatible with llama.cpp, enable_thinking is always injected? by TomTheWise · bug
#8688feat: built-in cpu-only llama cpp integration by tjbck · enhancement, help wanted
#17428issue: Support Think Parsing with llama.cpp + GPT-OSS by AbdullahMPrograms · bug
#17350issue: Llama.cpp server timing metrics not parsed correctly by ITankForCAD · bug
#16251issue: When using llama.cpp as backend, pressing stop doesn't stop token generation by OracleToes · bug
💡 If this is a duplicate, consider closing it and adding details to the existing issue.
This comment was generated automatically. React with 👍 if helpful, 👎 if not.
<!-- gh-comment-id:4368912033 -->
@owui-terminator[bot] commented on GitHub (May 4, 2026):
🔍 **Similar Issues Found**
I found some existing issues that might be related. Please check if any of these are duplicates or contain helpful solutions:
1. [#23703](https://github.com/open-webui/open-webui/issues/23703) **issue: Notes feature not compatible with llama.cpp, enable_thinking is always injected?**
*by TomTheWise · `bug`*
2. [#8688](https://github.com/open-webui/open-webui/issues/8688) **feat: built-in cpu-only llama cpp integration**
*by tjbck · `enhancement`, `help wanted`*
3. [#17428](https://github.com/open-webui/open-webui/issues/17428) **issue: Support Think Parsing with llama.cpp + GPT-OSS**
*by AbdullahMPrograms · `bug`*
4. [#17350](https://github.com/open-webui/open-webui/issues/17350) **issue: Llama.cpp server timing metrics not parsed correctly**
*by ITankForCAD · `bug`*
5. [#16251](https://github.com/open-webui/open-webui/issues/16251) **issue: When using llama.cpp as backend, pressing stop doesn't stop token generation**
*by OracleToes · `bug`*
---
💡 If this is a duplicate, consider closing it and adding details to the existing issue.
*This comment was generated automatically.* React with 👍 if helpful, 👎 if not.
<!-- gh-comment-id:4369764645 -->
@Classic298 commented on GitHub (May 4, 2026):
this can be easily done with a filter, did you look into it? https://docs.openwebui.com/features/extensibility/plugin/functions/filter
@Classic298 fair enough, though i do feel like it'd make sense as a builtin feature, we can already add custom args in the current chat, why couldn't we add them globaly in here?
<!-- gh-comment-id:4369926905 -->
@alkeryn commented on GitHub (May 4, 2026):
@Classic298 fair enough, though i do feel like it'd make sense as a builtin feature, we can already add custom args in the current chat, why couldn't we add them globaly in here?
<img width="1383" height="798" alt="Image" src="https://github.com/user-attachments/assets/cbe753ff-6c26-4a34-9e6a-ebc348562ec8" />
@alkeryn you can, scroll further down - you can add custom parameters. Open WebUI cannot had dozens and dozens more parameters for model or inference dependent engines, that's why you can add custom paramaters at the very end as well
<!-- gh-comment-id:4369946176 -->
@Classic298 commented on GitHub (May 4, 2026):
@alkeryn you can, scroll further down - you can add custom parameters. Open WebUI cannot had dozens and dozens more parameters for model or inference dependent engines, that's why you can add custom paramaters at the very end as well
but the "add new custom setting" doesn't show up in the global config here :
<!-- gh-comment-id:4370669568 -->
@alkeryn commented on GitHub (May 4, 2026):
@Classic298 you can only for a single chat ie :
<img width="428" height="1405" alt="Image" src="https://github.com/user-attachments/assets/36dc5683-d4f1-486f-9ff2-c91a249e9b20" />
but the "add new custom setting" doesn't show up in the global config here :
<img width="1682" height="938" alt="Image" src="https://github.com/user-attachments/assets/35a5242e-91ac-438f-a8fd-3ad4f3613d17" />
@alkeryn go to admin panel > models and add it to the model if your desire there in the advanced paramaters
<!-- gh-comment-id:4370731158 -->
@Classic298 commented on GitHub (May 4, 2026):
@alkeryn go to admin panel > models and add it to the model if your desire there in the advanced paramaters
@Classic298 sure, but per model is annoying, would be nice to have a global setting or at least per connection / provider.
<!-- gh-comment-id:4371671170 -->
@alkeryn commented on GitHub (May 4, 2026):
@Classic298 sure, but per model is annoying, would be nice to have a global setting or at least per connection / provider.
@alkeryn admin panel > settings > models > top right: settings - this should also do it
<!-- gh-comment-id:4371677021 -->
@Classic298 commented on GitHub (May 4, 2026):
@alkeryn admin panel > settings > models > top right: settings - this should also do it
<!-- gh-comment-id:4374562150 -->
@alkeryn commented on GitHub (May 4, 2026):
@Classic298 i don't see an option for it, weird.
<img width="1284" height="839" alt="Image" src="https://github.com/user-attachments/assets/c0433d51-ef01-406d-89bd-f34ab199d037" />
hmmmmmm open a feature request specifically for adding custom param IN THIS MODAL that you here opened so that you can add the same custom param to all models pls thx
<!-- gh-comment-id:4374582365 -->
@Classic298 commented on GitHub (May 4, 2026):
hmmmmmm open a feature request specifically for adding custom param IN THIS MODAL that you here opened so that you can add the same custom param to all models pls thx
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Originally created by @alkeryn on GitHub (May 4, 2026).
Original GitHub issue: https://github.com/open-webui/open-webui/issues/24341
Check Existing Issues
Verify Feature Scope
Problem Description
with llama.cpp it's possible to disable thinking through the api using : thinking_budget_tokens: 0, there may also be other methods that work with it.
anyway, currently there is no convenient builtin way to disable thinking for llama.cpp
it can be done per chat by adding "thinking_budget_tokens: 0" as a custom field in the chat settings, but there is no way to do it globaly.
Desired Solution you'd like
either add an option for llama.cpp in settings or allow custom fields not only in the chat settings but globaly.
Alternatives Considered
No response
Additional Context
No response
@owui-terminator[bot] commented on GitHub (May 4, 2026):
🔍 Similar Issues Found
I found some existing issues that might be related. Please check if any of these are duplicates or contain helpful solutions:
#23703 issue: Notes feature not compatible with llama.cpp, enable_thinking is always injected?
by TomTheWise ·
bug#8688 feat: built-in cpu-only llama cpp integration
by tjbck ·
enhancement,help wanted#17428 issue: Support Think Parsing with llama.cpp + GPT-OSS
by AbdullahMPrograms ·
bug#17350 issue: Llama.cpp server timing metrics not parsed correctly
by ITankForCAD ·
bug#16251 issue: When using llama.cpp as backend, pressing stop doesn't stop token generation
by OracleToes ·
bug💡 If this is a duplicate, consider closing it and adding details to the existing issue.
This comment was generated automatically. React with 👍 if helpful, 👎 if not.
@Classic298 commented on GitHub (May 4, 2026):
Filter?
@alkeryn commented on GitHub (May 4, 2026):
@Classic298 filter?
@Classic298 commented on GitHub (May 4, 2026):
this can be easily done with a filter, did you look into it? https://docs.openwebui.com/features/extensibility/plugin/functions/filter
@alkeryn commented on GitHub (May 4, 2026):
@Classic298 fair enough, though i do feel like it'd make sense as a builtin feature, we can already add custom args in the current chat, why couldn't we add them globaly in here?
@Classic298 commented on GitHub (May 4, 2026):
@alkeryn you can, scroll further down - you can add custom parameters. Open WebUI cannot had dozens and dozens more parameters for model or inference dependent engines, that's why you can add custom paramaters at the very end as well
@alkeryn commented on GitHub (May 4, 2026):
@Classic298 you can only for a single chat ie :
but the "add new custom setting" doesn't show up in the global config here :
@Classic298 commented on GitHub (May 4, 2026):
@alkeryn go to admin panel > models and add it to the model if your desire there in the advanced paramaters
@alkeryn commented on GitHub (May 4, 2026):
@Classic298 sure, but per model is annoying, would be nice to have a global setting or at least per connection / provider.
@Classic298 commented on GitHub (May 4, 2026):
@alkeryn admin panel > settings > models > top right: settings - this should also do it
@alkeryn commented on GitHub (May 4, 2026):
@Classic298 i don't see an option for it, weird.
@Classic298 commented on GitHub (May 4, 2026):
hmmmmmm open a feature request specifically for adding custom param IN THIS MODAL that you here opened so that you can add the same custom param to all models pls thx