ollama - ✅(Solved) Fix Can't disable thinking in any GUI, using Ollama 0.20.6 [3 pull requests, 4 comments, 4 participants]

Official PRs (…)
ON THIS PAGE

Recommended Tools

×6

Utilities matched from this issue’s tags and category — try them while you read without losing context.

GitHub issue graph ai analysis

Paste a GitHub issue URL. We fetch that issue, discover linked issues from bodies/comments/timeline, collect linked pull requests, and produce a structured English report.

The report is written in English Markdown for sharing and archival.

Helpful · Quick feedback

Loading…
GitHub stats
ollama/ollama#15536Fetched 2026-04-15 06:20:26
View on GitHub
Comments
4
Participants
4
Timeline
25
Reactions
3
Author
Timeline (top)
referenced ×5subscribed ×5commented ×4cross-referenced ×4

Fix Action

Fixed

PR fix notes

PR #15553: gemma4: fix nothink case renderer

Description (problem / solution / changelog)

Regressed in #15490

Fixes: #15536

Changed files

  • model/renderers/gemma4.go (modified, +2/-1)

PR #15556: Revert "gemma4: fix nothink case renderer"

Description (problem / solution / changelog)

Reverts ollama/ollama#15553

See https://github.com/ollama/ollama/issues/15536#issuecomment-4238939549

Changed files

  • model/renderers/gemma4.go (modified, +1/-2)

PR #15560: gemma4: restore e2b-style nothink prompt

Description (problem / solution / changelog)

Gemma 4 prompts differ when thinking is disabled for different sized models: 26b/31b emit an empty thought block, while e2b/e4b do not.

Before #15490, our shared Gemma 4 renderer effectively matched the e2b behavior. #15490 changed it to always emit the empty thought block, which regressed e2b/e4b nothink behavior and led to #15536 (and possibly

This change restores the previous shared behavior by removing the empty trailing thought block. It also renames the checked-in upstream chat templates so the e2b and 31b fixtures are tracked separately.

A follow-up will split Gemma 4 rendering by model size.

Fixes: #15536

Changed files

  • model/renderers/gemma4.go (modified, +0/-3)
  • model/renderers/gemma4_reference_test.go (modified, +85/-44)
  • model/renderers/testdata/gemma4_31b_chat_template.jinja2 (renamed, +0/-0)
  • model/renderers/testdata/gemma4_e2b_chat_template.jinja2 (added, +344/-0)
RAW_BUFFERClick to expand / collapse

What is the issue?

Since I updated to 0.20.6, I haven't been able to use "/set nothink" in the terminal, or any GUI with the ability to disable thinking. The model just ignores those settings, and outputs "thinking" reasoning anyway. Once I downgraded to 0.20.5, I can disable it fine again.

extent analysis

TL;DR

Downgrade to version 0.20.5 as a temporary workaround to resolve the issue with disabling "thinking" settings.

Guidance

  • The issue seems to be specific to version 0.20.6, as downgrading to 0.20.5 resolves the problem.
  • Verify that the "thinking" settings are correctly applied in version 0.20.5 to confirm the issue is version-specific.
  • Check the release notes or documentation for version 0.20.6 to see if there are any known issues or changes related to the "thinking" settings.
  • If possible, test other versions between 0.20.5 and 0.20.6 to identify the exact version where the issue was introduced.

Notes

The root cause of the issue is unclear, but it appears to be related to changes introduced in version 0.20.6.

Recommendation

Apply workaround: Downgrade to version 0.20.5, as it is the only known version where the "thinking" settings can be disabled correctly.

Vote matrix · Quick signals

Works
Did the solution work? Tap to confirm.
Easy Fix
Was it a quick fix?
Time Saver
Did it save you time?
Blocking
Was it severely blocking?
Common Issue
Are others likely hitting this too?
Flaky / Intermittent
Is it intermittent?
Verified / Reproducible
Can you reproduce it reliably?
Loading…

Still need to ship something?

×6

Another batch ranked right after the header list — different links, same matching logic.

Back to top recommendations

TRENDING