github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 16:40:08 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #10961] Enable / disable thinking through modelfiles and/or /save functionality. #69281

New Issue
Open
opened 2026-05-04 17:38:46 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-05-04 17:38:46 -05:00
Owner
Copy Link

Originally created by @stephensrmmartin on GitHub (Jun 3, 2025).
Original GitHub issue: https://github.com/ollama/ollama/issues/10961

We would like to serve models with the thinking setting preconfigured to either /think or /nothink .

Please allow us to change this via the modelfile, so that one can have, e.g., qwen3-think and qwen3-nothink variants preconfigured for users. They would be the same thing, ideally with one modelfile having THINK and another saying NOTHINK near the parameter section.

Moreover, this should be /save-able.

Originally created by @stephensrmmartin on GitHub (Jun 3, 2025). Original GitHub issue: https://github.com/ollama/ollama/issues/10961 We would like to serve models with the thinking setting preconfigured to either /think or /nothink . Please allow us to change this via the modelfile, so that one can have, e.g., qwen3-think and qwen3-nothink variants preconfigured for users. They would be the same thing, ideally with one modelfile having THINK and another saying NOTHINK near the parameter section. Moreover, this should be /save-able.
GiteaMirror added the thinkingfeature request labels 2026-05-04 17:38:46 -05:00
GiteaMirror commented 2026-05-04 17:38:48 -05:00
Author
Owner
Copy Link

@kanadrome commented on GitHub (Mar 2, 2026):

I tried 'PARAMETER think false' in a modelfile for Ollama 0.17.5 with the new Qwen 3.5:4b model and the ollama create command said

" unknown parameter 'think' "

I'm on MacOS. Should it not work in 0.17.5? Is the parser not updated yet? Or is it a different parameter for Qwen 3.5?

<!-- gh-comment-id:3987105980 --> @kanadrome commented on GitHub (Mar 2, 2026): I tried 'PARAMETER think false' in a modelfile for Ollama 0.17.5 with the new Qwen 3.5:4b model and the ollama create command said " unknown parameter 'think' " I'm on MacOS. Should it not work in 0.17.5? Is the parser not updated yet? Or is it a different parameter for Qwen 3.5?
GiteaMirror commented 2026-05-04 17:38:49 -05:00
Author
Owner
Copy Link

@randomkwiz commented on GitHub (Apr 6, 2026):

Please do it

<!-- gh-comment-id:4193517147 --> @randomkwiz commented on GitHub (Apr 6, 2026): Please do it
GiteaMirror commented 2026-05-04 17:38:50 -05:00
Author
Owner
Copy Link

@DarrenCook commented on GitHub (Apr 7, 2026):

I don't code in Go and haven't tested it myself yet (*), but https://github.com/ollama/ollama/pull/14108 looks quite small and straightforward. Can a developer look at it and get it merged in?

Being able to predefine qwen3-think and qwen3-nothink (and the same for gemma4 now) will be a big jump in ollama usefulness.

*: I'm running the quick install binary(?) version, so need to find out how involved it is to run directly from source.

<!-- gh-comment-id:4198952483 --> @DarrenCook commented on GitHub (Apr 7, 2026): I don't code in Go and haven't tested it myself yet (*), but https://github.com/ollama/ollama/pull/14108 looks quite small and straightforward. Can a developer look at it and get it merged in? Being able to predefine qwen3-think and qwen3-nothink (and the same for gemma4 now) will be a big jump in ollama usefulness. *: I'm running the quick install binary(?) version, so need to find out how involved it is to run directly from source.
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 thinking
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#69281
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?