feat: per-model thinkingDefault config and /think default directive#20458
Open
kmixter wants to merge 1 commit intoopenclaw:mainfrom
Open
feat: per-model thinkingDefault config and /think default directive#20458kmixter wants to merge 1 commit intoopenclaw:mainfrom
kmixter wants to merge 1 commit intoopenclaw:mainfrom
Conversation
kmixter
force-pushed
the
jasmine/per-model-thinking-default
branch
from
04057a8 to
82292ae
Compare
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
Comment on lines
166
to
174
| // Re-resolve thinking default if the model changed (per-model thinkingDefault support). | ||
| const newProvider = sessionEntry.providerOverride ?? provider; | ||
| const newModel = sessionEntry.modelOverride ?? model; | ||
| if (newProvider !== provider || newModel !== model) { | ||
| const catalog = await loadModelCatalog({ config: cfg }); | ||
| resolvedDefaultThinkLevel = | ||
| (sessionEntry?.thinkingLevel as ThinkLevel | undefined) ?? | ||
| resolveThinkingDefault({ cfg, provider: newProvider, model: newModel, catalog }); | ||
| } |
Contributor
There was a problem hiding this comment.
per-model thinking default only re-resolved in directive-only path, not when directives are combined with message content
the re-resolution logic here only applies when the user sends a directive-only message (e.g., just /model openai/gpt-4). when directives are combined with content (e.g., /model openai/gpt-4 what is 2+2?), the code takes the hasAnyDirective path (line 217), which uses applyInlineDirectivesFastLane. that path calls resolveCurrentDirectiveLevels which invokes modelState.resolveDefaultThinkingLevel(), but this closure was created with the original provider/model values and won't reflect the per-model default for the newly switched model.
this means users who switch models inline with their message won't get the correct per-model thinking default until their next message.
consider moving the re-resolution logic to a shared location that both paths use, or updating the resolveDefaultThinkingLevel closure to check for provider/model overrides in the session entry.
Prompt To Fix With AI
This is a comment left during a code review.
Path: src/auto-reply/reply/get-reply-directives-apply.ts
Line: 166-174
Comment:
per-model thinking default only re-resolved in directive-only path, not when directives are combined with message content
the re-resolution logic here only applies when the user sends a directive-only message (e.g., just `/model openai/gpt-4`). when directives are combined with content (e.g., `/model openai/gpt-4 what is 2+2?`), the code takes the `hasAnyDirective` path (line 217), which uses `applyInlineDirectivesFastLane`. that path calls `resolveCurrentDirectiveLevels` which invokes `modelState.resolveDefaultThinkingLevel()`, but this closure was created with the *original* provider/model values and won't reflect the per-model default for the newly switched model.
this means users who switch models inline with their message won't get the correct per-model thinking default until their next message.
consider moving the re-resolution logic to a shared location that both paths use, or updating the `resolveDefaultThinkingLevel` closure to check for provider/model overrides in the session entry.
How can I resolve this? If you propose a fix, please make it concise.
kmixter
force-pushed
the
jasmine/per-model-thinking-default
branch
from
82292ae to
493652c
Compare
6 tasks
Add per-model thinkingDefault to resolve thinking level based on which model is active, so users running different models (e.g. gemini-3-pro for chat, gemini-3-flash for heartbeats) get appropriate thinking levels automatically without manually toggling /think per session. Resolution priority: per-model → global → catalog auto-detect (existing fallback behavior unchanged). Config keys are normalized via parseModelRef so aliases like "anthropic/opus-4.6" resolve correctly. Also adds: - /think default directive to undo a session-level thinking override and cascade back to the per-model default - Re-resolve thinking default after inline /model switch for both directive-only and inline-with-content paths - Memory flush passes provider/model through for correct resolution Related: openclaw#18152 (built independently, shares the per-model config concept but adds /think default and model-switch re-resolution) AI-assisted (Claude Code), fully tested. Tested: pnpm build, pnpm check, pnpm test, and e2e verified on live gateway.
kmixter
force-pushed
the
jasmine/per-model-thinking-default
branch
from
493652c to
240f2b6
Compare
Author
|
The concern raised by the Greptile bot (per-model thinking default only re-resolved in directive-only path) has been addressed in the current revision:
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Add per-model thinkingDefault to resolve thinking level based on which
model is active, so users running different models (e.g. gemini-3-pro
for chat, gemini-3-flash for heartbeats) get appropriate thinking levels
automatically without manually toggling /think per session.
Resolution priority: per-model → global → catalog auto-detect (existing
fallback behavior unchanged). Config keys are normalized via parseModelRef
so aliases like "anthropic/opus-4.6" resolve correctly.
Also adds:
cascade back to the per-model default
directive-only and inline-with-content paths
Discussion: #20612
Related: #18152 (built independently, shares the per-model config concept
but adds /think default and model-switch re-resolution)
Config example
{ "agents": { "defaults": { "thinkingDefault": "low", "models": { "google/gemini-2.5-pro": {}, "google/gemini-3-flash": { "thinkingDefault": "high" }, "google/gemini-flash-lite-latest": { "thinkingDefault": "off" } } } } }Test plan
pnpm buildpassespnpm checkpasses (clean on changed files)pnpm testpasses (456 tests, 47 files)/think defaultresets level and per-model config resolves correctly viaopenclaw doctorAI-assisted (Claude Code), fully tested.