Using the q= URL parameters defaults the model to GPT4o even if you explicitly pass a model via the url using model=

ChatGPT Bugs
1

Hello, here’s a friendly bug report, I hope your PMs and Engineers will appreciate the format. :slight_smile:

Bug Report

Title: Using the q= URL parameter defaults the model to GPT4o, even when an explicit model= parameter is passed.

Description: When constructing a URL with the q= parameter to pre-fill a query, the system defaults the model to GPT4o regardless of whether a specific model is explicitly defined in the model= parameter. This behavior is unexpected and prevents users from controlling which model is used for their query.

Steps to Reproduce:

  1. Construct a URL that includes both q= and model= parameters, such as:
https://example.com/?q=test_query&model=o1
  1. Load the URL in a browser or relevant environment.
  2. Observe the model selection in the system interface.

Expected Behavior: The model specified in the model= parameter (e.g., o1 ) should be used for processing the query.

Actual Behavior: The system defaults to GPT4o regardless of the value passed in the model= parameter.

Impact: This bug limits the flexibility of URL-based query configuration and forces users into a default model (GPT4o) even when other models are explicitly desired.

Environment:

  • Platform: Web
  • Browser: [Specify browser and version]
  • Date/Time of Occurrence: [Specify date/time]
  • URL Example: https://example.com/?q=test_query&model=o1

Additional Information:

  • This issue may be related to how the q= parameter overrides or interacts with the model= parameter in the backend logic.
  • A workaround is not apparent since the q= parameter is essential for pre-filling queries.

Suggested Fix: Review and adjust the URL parsing logic to ensure that the model= parameter takes precedence over the default GPT4o when both parameters are provided.

2

Bumping this thread to see if we can get a response on this