Agent Security Evaluation on ASB (Agent Security Benchmark)
7ASR-d (ASB)gpt-4o + GBT-SE
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| gpt-4o + GBT-SEBase Model=gpt-4o, Safety/Security Framework=GBT-SE2026.01 | 7 | 78.4 | 70.4 | |
| llama-3-8b + GBT-SEBase Model=llama-3-8b, Safety/Security Framework=GBT-SE2026.01 | 7.4 | 77 | 51.8 | |
| gpt-4o + GBT-BasicBase Model=gpt-4o, Safety/Security Framework=GBT-Basic2026.01 | 8.6 | 73.8 | 70.1 | |
| llama-3-8b + GBT-BasicBase Model=llama-3-8b, Safety/Security Framework=GBT-Basic2026.01 | 8.7 | 72.2 | 51.6 | |
| llama-3-8b + Global guardrail onlyBase Model=llama-3-8b, Safety/Security Framework=Global guardrail only2026.01 | 10.9 | 61.6 | 50.8 | |
| gpt-4o + Global guardrail onlyBase Model=gpt-4o, Safety/Security Framework=Global guardrail only2026.01 | 13.1 | 63.2 | 69 | |
| llama-3-8b (native)Base Model=llama-3-8b, Safety/Security Framework=native2026.01 | 20.4 | 4.9 | 52 | |
| gpt-4o (native)Base Model=gpt-4o, Safety/Security Framework=native2026.01 | 64.4 | 8.8 | 71.3 |