General Reasoning on Reasoning Benchmark Suite (GSM8K, MATH500, GPQA, CSQA, AQuA, MMLU)
86.61Average AccuracyWeighted MV
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Weighted MV2026.04 | 86.61 | 9 | |
| Simple MV2026.04 | 86.38 | 9 | |
| EMS-SimSorting Strategy=Semantic Similarity2026.04 | 86.38 | 6.19 | |
| EMS-RelSorting Strategy=Historical Reliability2026.04 | 86.38 | 6.1 | |
| Conf-MV2026.04 | 85.79 | 9 | |
| Self-Consistency2026.04 | 84.91 | 10 | |
| Single Agent2026.04 | 83.92 | 1 |