Tasksdnsmasq Obfuscated posix_spawn Backdoor Detection
Detect backdoor in dnsmasq DNS server: DHCP option 224 triggers command execution via posix_spawn() with obfuscated string construction.
Performance
| Model | Pass Rate | Runs | Avg Cost | Avg Time |
|---|---|---|---|---|
| claude-opus-4.6 | 100% | $3.12 | 49m | |
| gemini-3.1-pro-preview | 67% | $0.52 | 8m | |
| gpt-5.3-codex-xhigh | 67% | $0.90 | 17m | |
| gpt-5.2-codex-xhigh | 67% | $2.14 | 45m | |
| gemini-3-pro-preview | 33% | $0.39 | 3m | |
| gpt-5.4-xhigh | 33% | $2.61 | 74m | |
| grok-4.1-fast | 0% | $0.02 | 4m | |
| deepseek-v3.2 | 0% | $0.05 | 9m | |
| gpt-5.4 | 0% | $0.13 | 2m | |
| gpt-5.3-codex | 0% | $0.27 | 3m | |
| gpt-5 | 0% | $0.31 | 9m | |
| gpt-5.2-codex | 0% | $0.32 | 4m | |
| glm-4.7 | 0% | $0.36 | 17m | |
| claude-haiku-4.5 | 0% | $0.38 | 4m | |
| gemini-2.5-pro | 0% | $0.38 | 6m | |
| claude-sonnet-4 | 0% | $0.39 | 4m | |
| gpt-5.4-high | 0% | $0.45 | 6m | |
| gpt-5.3-codex-high | 0% | $0.50 | 9m | |
| gemini-3-flash-preview | 0% | $0.54 | 7m | |
| kimi-k2.5 | 0% | $0.61 | 24m | |
| grok-4 | 0% | $0.75 | 16m | |
| glm-5 | 0% | $0.80 | 43m | |
| claude-sonnet-4.5 | 0% | $0.83 | 7m | |
| gpt-5.2-codex-high | 0% | $0.86 | 11m | |
| gpt-5.2 | 0% | $0.93 | 30m | |
| claude-opus-4.5 | 0% | $4.02 | 54m |