llamaIndex - 💡(How to fix) Fix [Feature Request]: Add support for GPT-5.4 and GPT-5.4-pro in OpenAIpackage [7 comments, 4 participants]

Fix: Update OpenAI LLMS Package to Support New Models

Step-by-Step Solution Plan

1. Update OpenAI LLMS Package: Clone the OpenAI LLMS package repository from GitHub.
2. Add Support for New Models: In the utils.py file, update the MODEL_VERSIONS dictionary to include the new model versions (gpt-5.4 and gpt-5.4-pro).
3. Update Model Loading Logic: Modify the load_model function to handle the new model versions.

Example Code:

# utils.py (updated)
MODEL_VERSIONS = {
    'gpt-3.5-turbo': 'text-davinci-003',
    'gpt-4': 'text-davinci-002',
    'gpt-5.4': 'text-davinci-004',  # New model version
    'gpt-5.4-pro': 'text-davinci-005',  # New model version
}

def load_model(model_name):
    if model_name in MODEL_VERSIONS:
        model_version = MODEL_VERSIONS[model_name]
        # Load the model using the Azure OpenAI package
        model = AzureOpenAI.load_model(model_version)
        return model
    else:
        raise ValueError(f"Unsupported model version: {model_name}")

1. Test the Updated Package: Run tests to ensure the updated package supports the new model versions.

Verification

• Test the load_model function with the new model versions (gpt-5.4 and gpt-5.4-pro).
• Verify that the model is loaded correctly and responds as expected.

Extra Tips

• Make sure to update the AzureOpenAI package to the latest version to ensure compatibility with the new model versions.
• Consider submitting a pull request to the OpenAI LLMS package repository to