Task-Oriented Dialogue on MultiWOZ 2.2 (test)
96.48Inform RateDarwinTOD
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| DarwinTODBackbone=GPT-5.12026.01 | 96.48 | 90.12 | 21.98 | 115.28 | |
| DarwinTODYear=2026, Backbone=GPT-5.12026.01 | 96.48 | 90.12 | 21.98 | 115.28 | |
| DarwinTODBackbone=GPT-42026.01 | 96.14 | 89.63 | 21.87 | 114.76 | |
| DarwinTODBackbone=Qwen3-14B2026.01 | 95.21 | 87.98 | 18.92 | 110.52 | |
| DarwinTODBackbone=Qwen3-8B2026.01 | 94.73 | 87.25 | 18.41 | 109.4 | |
| DarwinTODYear=2026, Backbone=Qwen3-8B2026.01 | 94.73 | 87.25 | 18.41 | 109.4 | |
| DarwinTODBackbone=Qwen2.5-14B2026.01 | 93.02 | 85.41 | 18.76 | 107.98 | |
| DarwinTODBackbone=Llama3.1-8B2026.01 | 92.59 | 84.09 | 18.27 | 106.61 | |
| DarwinTODBackbone=Llama3-8B2026.01 | 92.58 | 83.97 | 17.98 | 106.26 | |
| DarwinTODYear=2026, Backbone=Llama3-8B2026.01 | 92.58 | 83.97 | 17.98 | 106.26 | |
| DarwinTODBackbone=Qwen2.5-7B2026.01 | 92.14 | 84.33 | 18.34 | 106.58 | |
| DarwinTODYear=2026, Backbone=Qwen2.5-7B2026.01 | 92.14 | 84.33 | 18.34 | 106.58 | |
| DarwinTODBackbone=Qwen3-4B2026.01 | 91.87 | 84.08 | 18.12 | 106.1 | |
| AgentTODYear=2025, Status=Prev. SOTA2026.01 | 90.9 | 81.25 | 18.49 | 104.56 | |
| AgentTODBackbone=Llama3-8B2026.01 | 90.71 | 81.26 | 18.02 | 104 | |
| TOATODYear=20232026.01 | 90 | 79.8 | 17 | 101.9 | |
| DarwinTODBackbone=Qwen2.5-3B2026.01 | 86.31 | 77.95 | 17.21 | 99.34 | |
| BORTYear=20222026.01 | 85.5 | 77.4 | 17.9 | 99.4 | |
| GALAXYYear=20222026.01 | 85.4 | 75.7 | 19.6 | 100.2 | |
| UBARYear=20212026.01 | 83.4 | 70.3 | 17.6 | 94.4 | |
| SOLOISTYear=20212026.01 | 82.3 | 72.4 | 13.6 | 90.9 | |
| SGP-TODYear=20232026.01 | 82 | 72.5 | 9.22 | 86.47 | |
| MinTLYear=20202026.01 | 73.7 | 65.4 | 19.4 | 89 |