Mathematical Reasoning on GSM Hard (Accuracy, Latency, SLO)
62AccuracyINFRAMIND
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| INFRAMINDLoad level=10 req/min2026.06 | 62 | 37 | 100 | |
| INFRAMINDLoad level=50 req/min2026.06 | 60 | 76 | 98.8 | |
| INFRAMINDLoad level=100 req/min2026.06 | 57.2 | 81 | 98.4 | |
| GPTSwarmLoad level=50 req/min2026.06 | 57 | 707 | 14.8 | |
| MASRouterLoad level=100 req/min2026.06 | 56.4 | 652 | 20.6 | |
| MASRouterLoad level=50 req/min2026.06 | 55 | 521 | 27.4 | |
| MASRouterLoad level=10 req/min2026.06 | 54.6 | 38 | 100 | |
| GPTSwarmLoad level=10 req/min2026.06 | 54 | 52 | 100 | |
| GPTSwarmLoad level=100 req/min2026.06 | 54 | 806 | 9 | |
| MoALoad level=100 req/min2026.06 | 50 | 525 | 11.2 | |
| MoALoad level=10 req/min2026.06 | 49.3 | 38 | 100 | |
| MoALoad level=50 req/min2026.06 | 49.1 | 474 | 16.8 |