Software Generation on SRDD
0.818Metric CUA-ChatDev (Qwen 2.5)
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| UA-ChatDev (Qwen 2.5)Paradigm=multi-agent, Backbone=Qwen 2.52026.07 | 0.818 | 0.859 | 0.927 | 0.649 | |
| UA-ChatDev (Gemma 2)Paradigm=multi-agent, Backbone=Gemma 22026.07 | 0.767 | 0.84 | 0.929 | 0.596 | |
| ChatDevParadigm=multi-agent2026.07 | 0.56 | 0.88 | 0.802 | 0.395 | |
| GPT-EngineerParadigm=single-agent2026.07 | 0.502 | 0.358 | 0.789 | 0.142 | |
| MetaGPTParadigm=multi-agent2026.07 | 0.483 | 0.415 | 0.76 | 0.152 |