Success Rate on AgentBench
34.1Success RateDPO – ADARUBRIC-DA
Evaluation results
| Method | Links | |
|---|---|---|
| DPO – ADARUBRIC-DABackbone=Qwen2.5-7B, Evaluation Protocol=DPO training2026.03 | 34.1 | |
| DPO – ADARUBRIC-GMBackbone=Qwen2.5-7B, Evaluation Protocol=DPO training2026.03 | 31.9 | |
| DPO – ADARUBRIC-WMBackbone=Qwen2.5-7B, Evaluation Protocol=DPO training2026.03 | 30.8 | |
| DPO – PrometheusBackbone=Qwen2.5-7B, Evaluation Protocol=DPO training2026.03 | 26.4 | |
| DPO – G-EvalBackbone=Qwen2.5-7B, Evaluation Protocol=DPO training2026.03 | 24.5 | |
| DPO – Random pairsBackbone=Qwen2.5-7B, Evaluation Protocol=DPO training2026.03 | 21.3 | |
| SFT – Success onlyBackbone=Qwen2.5-7B2026.03 | 20.1 | |
| BaseBackbone=Qwen2.5-7B, Evaluation Protocol=zero-shot2026.03 | 15.2 |