Multi-agent collaborative coding on Multi-AgentBench Coding
50.7Task Completion Score (TCS)GPT-5-mini
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-5-miniAverage Steps (Stp)=12.0, Input Tokens (Tok In)=67.1, Output Tokens (Tok Out)=33.0, Execution Time=10425.62026.05 | 50.7 | |
| Gemini-3-flashAverage Steps (Stp)=28.2, Input Tokens (Tok In)=203.6, Output Tokens (Tok Out)=30.6, Execution Time=1191.32026.05 | 46.4 | |
| DeepSeek-V3.2Average Steps (Stp)=22.8, Input Tokens (Tok In)=117.9, Output Tokens (Tok Out)=9.0, Execution Time=639.82026.05 | 33.5 | |
| Qwen3-235B-A22BAverage Steps (Stp)=25.5, Input Tokens (Tok In)=161.5, Output Tokens (Tok Out)=20.9, Execution Time=1253.92026.05 | 28.2 |