feat(azure-open-ai): update model YAMLs [bot]#1019
Conversation
|
/test-models |
Gateway test results
|
![cursor[bot]](https://avatars.githubusercontent.com/in/1210556?s=60&v=4){kind=small}
There was a problem hiding this comment.
Cursor Bugbot has reviewed your changes and found 2 potential issues.
❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.
Reviewed by Cursor Bugbot for commit 9fe012e. Configure here.
| - prompt_caching | ||
| - structured_output | ||
| - system_messages | ||
| - tool_choice |
There was a problem hiding this comment.
Contradictory tool_choice in features and removeParams
High Severity
tool_choice was added to features by this PR, but it already exists in removeParams. This is contradictory — the model advertises tool_choice support while simultaneously stripping that parameter from API requests. Every other model in the codebase has tool_choice in one list or the other, never both. The sibling model o3-deep-research-2025-06-26-ev3.yaml correctly has tool_choice only in removeParams with no corresponding feature entry.
Additional Locations (1)
Reviewed by Cursor Bugbot for commit 9fe012e. Configure here.
| removeParams: | ||
| - response_format | ||
| - "n" | ||
| - n |
There was a problem hiding this comment.
Unquoted n parsed as boolean in YAML 1.1
Medium Severity
The quoted "n" in removeParams was changed to bare n in three files. In YAML 1.1 (used by tools like yamllint/PyYAML), bare n is interpreted as boolean false, not the string "n". The codebase convention is to quote it — the ModelParamKey type expects the string "n". While the project's js-yaml v4 (YAML 1.2) handles it correctly, this breaks consistency and portability.
Additional Locations (2)
Reviewed by Cursor Bugbot for commit 9fe012e. Configure here.
1 participant
Auto-generated by poc-agent for provider
azure-open-ai.Note
Medium Risk
Broad config-only updates to Azure OpenAI model definitions (pricing/limits/status/modes) could affect model selection, availability filtering, and cost calculations if downstream consumers rely on these fields.
Overview
Refreshes the
providers/azure-open-aimodel catalog with serverless provisioning metadata added across most model YAMLs and multiple pricing/limits/source corrections.Updates model lifecycle and compatibility metadata: several entries are now marked
deprecated/retiredwithdeprecationDate/isDeprecated, some models switch or addsupportedModes(notably addingresponsesand changing certain models fromchattoresponses), and modalities are expanded for newer video/realtime/image models (e.g., adding image/video inputs or audio outputs).Reviewed by Cursor Bugbot for commit 9fe012e. Bugbot is set up for automated code reviews on this repo. Configure here.