Long-context language understanding on LongBench 16K context slice v2
57.89ScoreQwen3.5-9B
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Qwen3.5-9BIndexer Type=Distilled2026.06 | 57.89 | 2.63 | |
| Qwen3.5-9BIndexer Type=Dense2026.06 | 55.26 | — | |
| Qwen3.5-0.8BIndexer Type=Distilled2026.06 | 31.58 | 13.16 | |
| Qwen3.5-0.8BIndexer Type=Dense2026.06 | 18.42 | — |