Advanced Paste: Switch from gpt-4o to gpt-4o-mini

[DeeQay]

Description of the new feature / enhancement

Summary:
Switching from GPT-4o to GPT-4o Mini in the Advanced Paste feature can save 94% on input tokens and 16.7x cheaper output tokens, with minimal performance difference. MMLU benchmark: 82% (GPT-4o Mini) vs. 88.7% (GPT-4o).

---

Change:

• File: src/modules/AdvancedPaste/AdvancedPaste/Services/OpenAI/KernelService.cs
• Line: 19
• Update: Change ModelName from "gpt-4o" to "gpt-4o-mini"

protected override string ModelName => "gpt-4o-mini";

Benefits

Cost savings with negligible performance impact.

Supporting information

openAI pricing
GPT4o Mini benchmarks