LLM Decoding Throughput on NOSA-8B 32K context length
2,312.1Throughput (tok/s)
SparDA
Updated 1mo ago
Evaluation Results
| Method | Links | |
|---|---|---|
2026.06 | 2,312.1 | |
2026.06 | 2,145.5 | |
2026.06 | 2,106.6 | |
2026.06 | 1,861.6 | |
2026.06 | 1,579.9 | |
2026.06 | 1,244.9 | |
2026.06 | 947.1 | |
2026.06 | 873.2 | |
2026.06 | 795.8 | |
2026.06 | 488.1 | |
2026.06 | 478 | |
2026.06 | 439.7 | |
2026.06 | 433.2 | |
2026.06 | 356.5 | |
2026.06 | 350.7 | |
2026.06 | 269.9 | |
2026.06 | 252 | |
2026.06 | 244 | |
2026.06 | 241.3 | |
2026.06 | 214.7 | |
2026.06 | 205.9 | |
2026.06 | 166 | |
2026.06 | 120.3 | |
2026.06 | 82 |