Classification on Synthetic Customer Scenarios
30.6Accuracy Gain (Synthetic Customer)GPT 5
Evaluation results
| Method | Links | ||
|---|---|---|---|
| GPT 5Model Backbone=GPT 52026.03 | 30.6 | 27.1 | |
| Gemini-1.5-ProModel Backbone=Gemini-1.5-Pro2026.03 | 28.4 | 24.8 | |
| GPT 4.1Model Backbone=GPT 4.12026.03 | 27.7 | 26.5 | |
| GPT 4.1-miniModel Backbone=GPT 4.1-mini2026.03 | 21.3 | 17.5 | |
| Claude Sonnet 3Model Backbone=Claude Sonnet 32026.03 | 20 | 16 |