github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 00:22:43 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #12493] Add /set silentthinking option #34054

New Issue
Open
opened 2026-04-22 17:17:11 -05:00 by GiteaMirror · 8 comments
GiteaMirror commented 2026-04-22 17:17:11 -05:00
Owner
Copy Link

Originally created by @iplayfast on GitHub (Oct 3, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/12493

Sometimes you want the model to think, but you don't want to suffer through the verbal diarrhea. An enhancement would be, don't output any text that is thinking, and only output the final resulting text.

Originally created by @iplayfast on GitHub (Oct 3, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/12493 Sometimes you want the model to think, but you don't want to suffer through the verbal diarrhea. An enhancement would be, don't output any text that is thinking, and only output the final resulting text.
GiteaMirror added the feature request label 2026-04-22 17:17:11 -05:00
GiteaMirror commented 2026-04-22 17:17:13 -05:00
Author
Owner
Copy Link

@FieldMouse-AI commented on GitHub (Oct 5, 2025):

Sometimes you want the model to think, but you don't want to suffer through the verbal diarrhea. An enhancement would be, don't output any text that is thinking, and only output the final resulting text.

Hello, @iplayfast , am I to understand you correctly, based on your question it would suggest that thinking being on changes how a model produces its response?

That surprises me. I thought that it was just like a kind of verbose option that lets us see how the model produced a response without changing the actual response.

Can you help clarify this for me, please?

<!-- gh-comment-id:3368828273 --> @FieldMouse-AI commented on GitHub (Oct 5, 2025): > Sometimes you want the model to think, but you don't want to suffer through the verbal diarrhea. An enhancement would be, don't output any text that is thinking, and only output the final resulting text. Hello, @iplayfast , am I to understand you correctly, based on your question it would suggest that `thinking` being on changes how a model produces its response? That surprises me. I thought that it was just like a kind of **verbose** option that lets us see how the model produced a response without changing the actual response. Can you help clarify this for me, please?
GiteaMirror commented 2026-04-22 17:17:13 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Oct 5, 2025):

The tokens produced during inference influence the following tokens. Thinking is used by models to "reason" about an answer before generating it. Disabling thinking (on models that support it) removes those early tokens and the final result may be different.

<!-- gh-comment-id:3369210608 --> @rick-github commented on GitHub (Oct 5, 2025): The tokens produced during inference influence the following tokens. Thinking is used by models to "reason" about an answer before generating it. Disabling thinking (on models that support it) removes those early tokens and the final result may be different.
GiteaMirror commented 2026-04-22 17:17:13 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Oct 5, 2025):

In case OP is unaware, the ollama client has the --hidethinking command line argument, but it's not settable via /set, which is what I think the OP wants.

<!-- gh-comment-id:3369215333 --> @rick-github commented on GitHub (Oct 5, 2025): In case OP is unaware, the ollama client has the `--hidethinking` command line argument, but it's not settable via `/set`, which is what I think the OP wants.
GiteaMirror commented 2026-04-22 17:17:13 -05:00
Author
Owner
Copy Link

@iplayfast commented on GitHub (Oct 5, 2025):

I was not actually aware of the --hidethinking command. So I guess /set hidethinking is what I'm looking for.

<!-- gh-comment-id:3369340911 --> @iplayfast commented on GitHub (Oct 5, 2025): I was not actually aware of the --hidethinking command. So I guess /set hidethinking is what I'm looking for.
GiteaMirror commented 2026-04-22 17:17:14 -05:00
Author
Owner
Copy Link

@YuenSzeHong commented on GitHub (Jan 31, 2026):

i am on podman (docker) 0.15.2, with 5070ti, for some reason, in gpt-oss and qwen3 /set nothink didnt turn off thinking

<!-- gh-comment-id:3828385980 --> @YuenSzeHong commented on GitHub (Jan 31, 2026): i am on podman (docker) 0.15.2, with 5070ti, for some reason, in gpt-oss and qwen3 /set nothink didnt turn off thinking
GiteaMirror commented 2026-04-22 17:17:14 -05:00
Author
Owner
Copy Link

@rick-github commented on GitHub (Jan 31, 2026):

Thinking cannot be disabled for gpt-oss, just set to a different level: high, medium, low. Some qwen3 models come in two variants: instruct and thinking. If you want to disable thinking, use the non-thinking model, eg qwen3:4b-instruct.

<!-- gh-comment-id:3828539188 --> @rick-github commented on GitHub (Jan 31, 2026): Thinking cannot be disabled for gpt-oss, just set to a different level: high, medium, low. Some qwen3 models come in two variants: instruct and thinking. If you want to disable thinking, use the non-thinking model, eg qwen3:4b-instruct.
GiteaMirror commented 2026-04-22 17:17:15 -05:00
Author
Owner
Copy Link

@iplayfast commented on GitHub (Feb 3, 2026):

I thought it was basically filtering out the thinking and only showing the results, I don't really care if there is thinking going on under the surface, I just don't want to see it.

<!-- gh-comment-id:3843047175 --> @iplayfast commented on GitHub (Feb 3, 2026): I thought it was basically filtering out the thinking and only showing the results, I don't really care if there is thinking going on under the surface, I just don't want to see it.
GiteaMirror commented 2026-04-22 17:17:15 -05:00
Author
Owner
Copy Link

@AstroTheRabbit commented on GitHub (Mar 13, 2026):

Recently installed ollama, and I'd definitely like a hidethinking parameter. Being able to hide the thinking output by passing a command to the CLI itself, but not being able to set it permanently (in a modelfile PARAMETER) or being able to enable it mid-session (via /set ...) seems kinda weird.

<!-- gh-comment-id:4052743991 --> @AstroTheRabbit commented on GitHub (Mar 13, 2026): Recently installed ollama, and I'd definitely like a `hidethinking` parameter. Being able to hide the thinking output by passing a command to the CLI itself, but not being able to set it permanently (in a modelfile `PARAMETER`) or being able to enable it mid-session (via `/set ...`) seems kinda weird.
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label feature request
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#34054
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?