Switch default OpenAI model from gpt-4o to gpt-5.4-mini

[Copilot]

This updates the OpenAI-compatible inference default to gpt-5.4-mini instead of gpt-4o. OpenAI-backed flows now use the new model without requiring any user configuration changes.

• Default model update

  • Changed the OpenAI model constant in src/llm.rs:
    • OPENAI_MODEL = "gpt-4o" → OPENAI_MODEL = "gpt-5.4-mini"

• Behavior impact

  • Applies to requests that resolve to the OpenAI provider path (OPENAI_API_KEY present).
  • No provider-selection logic or request structure changed.

const OPENAI_BASE: &str = "https://api.openai.com/v1";
const OPENAI_MODEL: &str = "gpt-5.4-mini";

Original prompt

Change the default OpenAI model from gpt-4o to gpt-5.4-mini

Created from Copilot CLI via the copilot delegate command.