Mathematical Reasoning on GSM8K (Throughput, Speedup)
1,499.6Throughput
SpecBundle
Updated 1mo ago
Evaluation Results
| Method | Links | ||
|---|---|---|---|
2026.03 | 1,499.6 | 1.4 | |
2026.03 | 1,323 | 1.96 | |
2026.03 | 1,071.3 | 1 | |
2026.03 | 971.9 | 2.13 | |
| 816.6 | 1.79 | ||
2026.03 | 758.3 | 1.62 | |
2026.03 | 722 | 1.59 | |
| 697 | 1.49 | ||
2026.03 | 674.3 | 1 | |
2026.03 | 544.2 | 1.61 | |
| 521.5 | 1.15 | ||
2026.03 | 469.1 | 1 | |
2026.03 | 455.9 | 1 | |
2026.03 | 453.2 | 1 | |
2026.03 | 337.9 | 1 | |
2026.03 | 329.7 | 1.81 | |
| 228.6 | 1.26 | ||
2026.03 | 181.8 | 1 |