fix: Google: use maxTokens from params #734
Merged
Conversation
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
15 tasks
nicolasf
force-pushed
the
fix-google-max-tokens-hardcoded
branch
from
d5fea78 to
174924f
Compare
Current implementation has a hardcoded maxOutputTokens=2048. Use the value from prompt.params.maxTokens similar to the OpenAILLMClient.
nicolasf
force-pushed
the
fix-google-max-tokens-hardcoded
branch
from
174924f to
bd14e68
Compare
|
Hi @kpavlov. |
|
HI @kpavlov. Is it possible to get this merged before 0.5.0 is released? |
|
Thanks for fixing this @nicolasf! @Ololoshechkin, are there plans to release 0.5.0 or 0.4.3 soon (or even a feature or SNAPSHOT build)? We're loving Koog so far, but we cannot migrate to it fully, as this issue prevents the use of Gemini models. |
Ololoshechkin
pushed a commit
that referenced
this pull request
Oct 2, 2025
(cherry picked from commit 45012a3)
karloti
pushed a commit
to karloti/koog
that referenced
this pull request
Oct 21, 2025
(cherry picked from commit 45012a3)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
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Outdated suggestions cannot be applied.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.
Current implementation has a hardcoded maxOutputTokens=2048. Use the value from prompt.params.maxTokens similar to the OpenAILLMClient.
Motivation and Context
During execution of structured output requests using Gemini 2.5 I encountered unexpected failures due to reaching MAX_TOKENS.
After checking the GoogleLLMClient implementation I noticed a hardcoded
maxOutputTokens = 2048.Checking the other clients, I'm replicating the same logic that exists on
OpenAILLMClient, which applies the value fromprompt.params.maxTokens.Breaking Changes
No breaking changes.
Type of the changes
Checklist
developas the base branchAdditional steps for pull requests adding a new feature