Mathematical Problem Solving on MATH (Accuracy, Latency, SLO)
82AccuracyINFRAMIND
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| INFRAMINDLoad level=10 req/min2026.06 | 82 | 44 | 100 | |
| INFRAMINDLoad level=50 req/min2026.06 | 75.6 | 255 | 87.3 | |
| INFRAMINDLoad level=100 req/min2026.06 | 75.2 | 290 | 61.8 | |
| MoALoad level=100 req/min2026.06 | 74.8 | 1,160 | 0.4 | |
| MoALoad level=10 req/min2026.06 | 74.4 | 122 | 100 | |
| MoALoad level=50 req/min2026.06 | 73.8 | 1,145 | 3.2 | |
| MASRouterLoad level=50 req/min2026.06 | 72.8 | 321 | 53 | |
| GPTSwarmLoad level=50 req/min2026.06 | 71 | 1,209 | 7.4 | |
| MASRouterLoad level=100 req/min2026.06 | 70.6 | 399 | 48.2 | |
| GPTSwarmLoad level=10 req/min2026.06 | 69.6 | 84 | 100 | |
| GPTSwarmLoad level=100 req/min2026.06 | 68.4 | 1,325 | 2.2 | |
| MASRouterLoad level=10 req/min2026.06 | 67.4 | 50 | 100 |