| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Math | Real-data Math Mistral outputs (3% verified data) | Original Accuracy50.46 | 6 | |
| Math | Real-data Math Qwen outputs (3% verified data) | Original Accuracy55.1 | 6 | |
| Math | Real-data Math Gemini outputs (3% verified data) | Original Accuracy66.35 | 6 | |
| Math | Real-data Math GPT-5.4-mini (3% verified data) | Original Accuracy62.35 | 6 | |
| Math | Real-data Math GPT-5.4 outputs (3% verified data) | Original Accuracy67.2 | 6 |