Long-context language understanding on LongBench 32K context slice v2
56.86ScoreQwen3.5-9B
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Qwen3.5-9BIndexer Type=Distilled2026.06 | 56.86 | 1.96 | |
| Qwen3.5-9BIndexer Type=Dense2026.06 | 54.9 | — | |
| Qwen3.5-0.8BIndexer Type=Dense2026.06 | 29.41 | — | |
| Qwen3.5-0.8BIndexer Type=Distilled2026.06 | 29.41 | 0 |