fix(llm): exclude sampling params for reasoning models#1822
Merged
Conversation
…s (o1, o3, o4) Bump uipath to 2.13.13. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
github-actions
Bot
added
test:uipath-langchain
Contributor
There was a problem hiding this comment.
Pull request overview
This PR updates the UiPath LLM Gateway client(s) to avoid sending OpenAI sampling parameters that cause 400s for OpenAI “reasoning” models (o1/o3/o4), and bumps the uipath package version accordingly.
Changes:
- Filter out
temperaturefor reasoning models inUiPathOpenAIService.chat_completions. - Filter out
temperature,n,frequency_penalty,presence_penalty, andtop_pfor reasoning models inUiPathLlmChatService.chat_completions. - Add tests covering reasoning-model filtering for the normalized LLM service; bump
uipathto2.13.13.
Reviewed changes
Copilot reviewed 3 out of 4 changed files in this pull request and generated 1 comment.
| File | Description |
|---|---|
| packages/uipath-platform/src/uipath/platform/chat/_llm_gateway_service.py | Adds reasoning-model detection and excludes unsupported params from request bodies. |
| packages/uipath-platform/tests/services/test_llm_service.py | Adds tests asserting reasoning models exclude sampling params (normalized API). |
| packages/uipath/pyproject.toml | Bumps uipath version to 2.13.13. |
| packages/uipath/uv.lock | Updates lockfile version entry for uipath to 2.13.13. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
…ests Covers the passthrough (non-normalized) API path that also conditionally omits temperature for o1/o3/o4 models. Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
|
Chibionos
reviewed
Jul 20, 2026
| # Build request body - some models don't support certain parameters | ||
| model_lower = model.lower() | ||
| is_claude_model = "claude" in model_lower | ||
| is_reasoning_model = model_lower.startswith(("o1", "o3", "o4")) |
Contributor
There was a problem hiding this comment.
using model name is unreliable sadly. But I think this is a best effort we can live with.
Chibionos
approved these changes
Jul 20, 2026
Chibionos
left a comment
Chibionos
left a comment
Contributor
There was a problem hiding this comment.
closest fix as BYOM can bring custom names too.
milind-jain-uipath
added a commit
that referenced
this pull request
Jul 21, 2026
0.2.11 was released to PyPI from main (#1822) without this change, so this PR ships as 0.2.12. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
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
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.
Jira
SRE-625200
Summary
temperature,n,frequency_penalty,presence_penalty, andtop_pfor OpenAI reasoning models (o1, o3, o4) which don't support these parameters and return 400 errors when they're included.UiPathOpenAIServiceandUiPathLlmChatService.uipathversion to 2.13.13.Failure Details (from SRE-625200)
Alert:
agents-eu-prd-agents-evalrun-failed-urt-[Agent Evaluations]-logalertv2-jaeDescription: Monitors
EvalRun.FailedandEvalSetRun.Failedevents with URT runtime whenErrorMessageis not null. The alert fired in thejae(Japan East) region on 2026-07-18.Alert details:
jae(Japan East)2026-07-18T06:12:26Zand2026-07-18T06:27:26ZRoot cause: OpenAI reasoning models (o1, o3, o4) do not support
temperatureand other sampling parameters (n,frequency_penalty,presence_penalty,top_p). Sending these parameters causes 400 errors, which surface asEvalRun.Failedevents.Search query that triggered the alert:
Test plan
🤖 Generated with Claude Code