feat: support gpt-5 for self-hosted model

What does this merge request do and why?

Self-hosted GitLab instances using AI Gateway encounter 500 errors when attempting to use openai/gpt-5 or openai/o3 models due to LiteLLM compatibility issues with newer OpenAI model requirements.

There aree a few issues with the following configuration of the gpt-5 model

• LiteLLM bug fix for O-series and GPT-5 parameter handling: OpenAI gpt-5 series does not support "max_tokens" parameter and temperature values that are not = 1 #

  • This fix addresses the max_tokens vs max_completion_tokens parameter issue and temperature constraints
  • Fixed in !3800 (merged)

• Updated chat/react/gpt/1.0.0.yml to use temperature: 1.0 for GPT ReAct Chat agent

  • There was no family for the following model_configuration check to use the base temperature 0.0 as a parameter which prevented using the following gpt-5 models.

• Removed unsupported stop parameter that may cause issues with newer models

  >> from litellm import get_supported_openai_params
  >>> 
  >>> def show_gpt5_params(model: str = "gpt-5") -> None:
  ...     """Display supported OpenAI parameters for GPT-5 models."""
  ...     params = get_supported_openai_params(model=model)
  ...     
  ...     print(f"\nSupported parameters for {model}:")
  ...     print("-" * 60)
  ...     for param in sorted(params):
  ...         print(f"  • {param}")
  ...     print(f"\nTotal: {len(params)} parameters")
  ... 
  >>> # Usage
  >>> show_gpt5_params("gpt-5")
  
  Supported parameters for gpt-5:
  ------------------------------------------------------------
    • audio
    • extra_headers
    • function_call
    • functions
    • logit_bias
    • max_completion_tokens
    • max_retries
    • max_tokens
    • modalities
    • n
    • parallel_tool_calls
    • prediction
    • reasoning_effort
    • response_format
    • safety_identifier
    • seed
    • service_tier
    • stream
    • stream_options
    • temperature
    • tool_choice
    • tools
    • user
    • web_search_options
  
  Total: 24 parameters

How to set up and validate locally

• Set up self-managed Gitlab instance with AI Gateway
• Navigate to GitLab Duo Self-Hosted in Duo admin dashboard
• Set up a self-hosted model pointing at https://api.openai.com/v1 with model identifier openai/gpt-5 or openai/o3 and include a valid OpenAI API key
• Click Test connection
• Change model identifier to openai/gpt-4o . Test connection will work successfully

Model Setup	Request

Merge request checklist

• Tests added for new functionality. If not, please raise an issue to follow up.
• Documentation added/updated, if needed.
• If this change requires executor implementation: verified that issues/MRs exist for both Go executor and Node executor or confirmed that changes are backward-compatible and don't break existing executor functionality.