svenstaro/genact
🌀 A nonsense activity generator
Results by Model
| # | Model | Score Percentage of hidden behavioral tests passed. | Cost Total API cost in USD for this task instance. | Calls Number of LLM API calls for this task instance. | |
|---|---|---|---|---|---|
| 1 |  |
Claude Sonnet 4.6 Anthropic | 59.1% | $17.93 | 298 |
| 2 | |
Claude Opus 4.6 Anthropic | 53.0% | $11.11 | 158 |
| 3 |  |
Gemini 3.1 Pro Google | 35.8% | $1.25 | 44 |
| 4 |  |
GPT 5.4 OpenAI | 24.1% | $0.46 | 10 |
| 5 | |
Claude Haiku 4.5 Anthropic | 24.1% | $0.58 | 79 |
| 6 | |
Claude Opus 4.7 Anthropic | 23.3% | $1.70 | 49 |
| 7 | |
Gemini 3 Flash Google | 20.7% | $0.47 | 48 |
| 8 | |
GPT 5.4 mini OpenAI | 12.5% | $0.02 | 10 |
| 9 | |
GPT 5 mini OpenAI | 1.7% | $0.02 | 9 |
Click row to see model details