Long-context evaluation on 128K context
80.12Quality Score (Q)Dense
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| DenseModel=GPT-oss2026.05 | 80.12 | — | 59.9 | — | 2,442.2 | — | 0.313 | |
| SphKV (rep)Model=GPT-oss2026.05 | 79.58 | 54 | 102.4 | 1.71 | 1,611.7 | 34 | 0.206 | |
| DenseModel=Qwen2.5-14B2026.05 | 78.02 | — | 55.7 | — | 2,775.6 | — | 0.355 | |
| SphKV (rep)Model=Qwen2.5-14B2026.05 | 77.4 | 62 | 94.8 | 1.7 | 1,607.7 | 42.1 | 0.206 | |
| DenseModel=Llama-3.1-8B2026.05 | 76.56 | — | 62.9 | — | 2,360.6 | — | 0.302 | |
| SphKV (rep)Model=Llama-3.1-8B2026.05 | 75.78 | 78 | 108 | 1.72 | 1,547.1 | 34.5 | 0.198 |