Database interaction on AgentBench DB
74.3Task Completion Score (TCS)Gemini-3-flash
Evaluation Results
| Method | Links | |
|---|---|---|
| Gemini-3-flashAverage Steps (Stp)=4.5, Input Tokens (Tok In)=14.9, Output Tokens (Tok Out)=5.3, Execution Time=60.02026.05 | 74.3 | |
| GPT-5-miniAverage Steps (Stp)=4.3, Input Tokens (Tok In)=6.5, Output Tokens (Tok Out)=2.3, Execution Time=30.82026.05 | 74 | |
| DeepSeek-V3.2Average Steps (Stp)=7.5, Input Tokens (Tok In)=15.9, Output Tokens (Tok Out)=0.7, Execution Time=30.22026.05 | 71.3 | |
| Qwen3-235B-A22BAverage Steps (Stp)=2.3, Input Tokens (Tok In)=3.0, Output Tokens (Tok Out)=0.1, Execution Time=21.52026.05 | 53.7 |