End-to-end Text Generation on NVIDIA B200 GPU Overall Generation
4,920.12Throughput (tokens/s)YOCO (CLSA)
Evaluation Results
| Method | Links | |
|---|---|---|
| YOCO (CLSA)Context Length=8K, Hardware=NVIDIA B200 GPU2026.06 | 4,920.12 | |
| YOCO (Dense)Context Length=8K, Hardware=NVIDIA B200 GPU2026.06 | 4,311.58 | |
| YOCO (CLSA)Context Length=16K, Hardware=NVIDIA B200 GPU2026.06 | 3,715.19 | |
| TransformerContext Length=8K, Hardware=NVIDIA B200 GPU2026.06 | 2,989.78 | |
| YOCO (Dense)Context Length=16K, Hardware=NVIDIA B200 GPU2026.06 | 2,904.96 | |
| YOCO (CLSA)Context Length=32K, Hardware=NVIDIA B200 GPU2026.06 | 2,805.48 | |
| YOCO (Dense)Context Length=32K, Hardware=NVIDIA B200 GPU2026.06 | 1,832.66 | |
| YOCO (CLSA)Context Length=64K, Hardware=NVIDIA B200 GPU2026.06 | 1,735.59 | |
| TransformerContext Length=16K, Hardware=NVIDIA B200 GPU2026.06 | 1,449.91 | |
| YOCO (CLSA)Context Length=128K, Hardware=NVIDIA B200 GPU2026.06 | 1,068.06 | |
| YOCO (Dense)Context Length=64K, Hardware=NVIDIA B200 GPU2026.06 | 1,042.9 | |
| YOCO (Dense)Context Length=128K, Hardware=NVIDIA B200 GPU2026.06 | 599.05 | |
| TransformerContext Length=32K, Hardware=NVIDIA B200 GPU2026.06 | 570.27 | |
| TransformerContext Length=64K, Hardware=NVIDIA B200 GPU2026.06 | 191.36 | |
| TransformerContext Length=128K, Hardware=NVIDIA B200 GPU2026.06 | 62.53 |