Wire · ai
FAR.AI Launches AI Security Leaderboard Revealing Hundredfold Gap in Frontier AI Model ...
◆ Sectors
◆ Source
◆ Verified
Fusion42 · 29 July 2026 · Fusion42 review
AI security research nonprofit FAR.AI has launched a public leaderboard ranking the security of frontier AI models. The key finding reveals a hundredfold cost gap to 'jailbreak' them: attacks on Grok 4.5 and Gemini 3.1 Pro succeeded for under $300, while identical methods failed against Claude Fable 5 and GPT-5.6 Sol after incurring costs of over $14,200.
This Wire brief sits within Fusion42's coverage of AI Frontier Models.
◆ ◆ The Wire takeaway
Your choice of foundational model is now a measurable security risk. This new leaderboard shows that breaking applications built on Gemini or Grok costs less than $300, a vulnerability you now own.
◆ Coverage
1 source · 29 Jul 2026
◆ Related on Wire
◆ Topics