OpenAI API: gpt-5-mini Support

Overview

This PR adds support for gpt-5-mini and newer gpt-5 models in the OpenAI API client. The changes address API parameter specification differences in the newer models.

Changes

1. API Parameter Changes

• Changed max_tokens → max_completion_tokens
  • Newer gpt-5 models require max_completion_tokens instead of max_tokens

2. Default Value Updates

• Default model: gpt-4-turbo-preview → gpt-5-mini
• Default temperature: 0.7 → 1.0
  • gpt-5 and newer models only support temperature value of 1

3. Configuration Comments Update

• Updated model selection comments
• Added note about gpt-5+ temperature restrictions

Modified Files

• src/main/java/com/steve/ai/ai/OpenAIClient.java
  • Changed to use max_completion_tokens parameter
• src/main/java/com/steve/ai/config/SteveConfig.java
  • Changed default model to gpt-5-mini
  • Changed default temperature to 1.0
  • Updated comments

Testing

• Verified API requests work correctly with gpt-5-mini model
• Confirmed error responses (unsupported_parameter, unsupported_value) are resolved

Compatibility

• Older models (gpt-3.5-turbo, gpt-4, etc.) will still work, but will use max_completion_tokens parameter
• You can still use older models by changing the model name or temperature value in the config file

Related Errors

The following errors are resolved:

• Unsupported parameter: 'max_tokens' is not supported with this model
• Unsupported value: 'temperature' does not support 0.7 with this model