Long-horizon agentic search on xbench DeepSearch
56.2CCGLM-5.0
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| GLM-5.02026.06 | 56.2 | 2.8 | 26.6 | 10 | 4.4 | 0 | |
| Qwen3.5-397B-A17B2026.06 | 55.2 | 1 | 23.4 | 14 | 6.4 | 0 | |
| GLM-4.72026.06 | 54.6 | 1.4 | 27 | 5.4 | 11.6 | 0 | |
| MiniMax-M2.52026.06 | 49.4 | 2 | 26.4 | 22.2 | 0 | 0 |