Safety Evaluation on AgentHarm
0.0001Cost per Accuracy Point ($)Llama 4 Maverick 17B
Evaluation Results
| Method | Links | |
|---|---|---|
| Llama 4 Maverick 17B2026.03 | 0.0001 | |
| Gemini 2.5 Pro2026.03 | 0.0007 | |
| GPT 4o2026.03 | 0.0013 | |
| Claude Sonnet 4.52026.03 | 0.005 |
| Method | Links | |
|---|---|---|
| Llama 4 Maverick 17B2026.03 | 0.0001 | |
| Gemini 2.5 Pro2026.03 | 0.0007 | |
| GPT 4o2026.03 | 0.0013 | |
| Claude Sonnet 4.52026.03 | 0.005 |