AIWiredabout 7 hours ago
Open AI Models Match Closed Ones in Hacking
4 min read
New research shows open-weight AI models match closed ones in hacking capability, with DeepSeek completing 73 percent of tasks versus 81 percent for OpenAI's Daybreak. Costs range from $2 to $46 per attack, challenging the strategy of restricting access to advanced models.
Level
Hype check
Tap to vote and see what everyone thinks.
