Detect backdoor in dnsmasq DNS server: DHCP option 224 triggers command execution via posix_spawn() instead of execl().
Performance
| Model | Pass Rate | Runs | Avg Cost | Avg Time |
|---|---|---|---|---|
| gemini-3-pro-preview | 100% | | $0.14 | 2m |
| gemini-3-flash-preview | 100% | | $0.16 | 3m |
| gemini-2.5-pro | 100% | | $0.19 | 3m |
| kimi-k2.5 | 67% | | $0.19 | 11m |
| glm-4.7 | 67% | | $0.22 | 9m |
| gpt-5.2 | 67% | | $0.38 | 15m |
| claude-opus-4.5 | 67% | | $2.48 | 37m |
| claude-opus-4.6 | 67% | | $5.18 | 62m |
| claude-sonnet-4 | 33% | | $0.50 | 5m |
| grok-4 | 33% | | $0.59 | 12m |
| grok-4.1-fast | 0% | | $0.01 | 2m |
| deepseek-v3.2 | 0% | | $0.04 | 8m |
| gpt-5.2-codex | 0% | | $0.23 | 3m |
| gpt-5 | 0% | | $0.39 | 11m |
| claude-haiku-4.5 | 0% | | $0.42 | 5m |
| claude-sonnet-4.5 | 0% | | $1.12 | 11m |
All product names, logos, and brands (™/®) are the property of their respective owners; they're used here solely for identification and comparison, and their use does not imply affiliation, endorsement, or sponsorship.